82530 JOP27710.1177/0269881113482530Journal of PsychopharmacologyShin et al.

Original Paper

Risk of ischemic stroke with the use of

risperidone, quetiapine and olanzapine
in elderly patients: a population-based, Journal of Psychopharmacology

case-crossover study 27(7) 638­–644

© The Author(s) 2013
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0269881113482530
jop.sagepub.com

Ju-Young Shin1, Nam-Kyong Choi2, Sun-Young Jung3,

Joongyub Lee2, Jun S Kwon4 and Byung-Joo Park1,2,3,5

Abstract
We conducted a case-crossover study to evaluate the comparative risk of ischemic stroke associated with the use of risperidone, quetiapine and
olanzapine in geriatric patients using the Korean Health Insurance Review and Assessment Service database. Cases included elderly patients >64 years
old who had experienced their first ischemic stroke (International Classification of Disease, Tenth Revision (ICD-10), I63) hospitalization from July
2005 to June 2006 and who had been without prior cerebrovascular diseases (ICD-10, I60–I69), or transient ischemic attack (ICD-10, G45). Exposures
to risperidone, quetiapine and olanzapine were assessed during the 30 days prior to the stroke episode. We set two control periods with lengths
which were the same as the hazard periods. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were estimated by conditional logistic
regression. A total of 1601 cases of ischemic stroke with a mean age of 75.6 (±6.7) years were identified, among which 933 (58.3%) were female. An
increased risk of ischemic stroke was associated with the use of risperidone (aOR=3.5, 95% CI 3.3–4.6) and quetiapine (aOR=2.7, 95% CI 2.0–3.6)
during the 30 days prior to stroke: however, no significant risk was observed with olanzapine (aOR=1.2, 95% CI 0.7–2.0). The increased stroke risk in
demented patients, assessed within 30 days after exposure, was also observed with olanzapine. However, the sample of olanzapine users was small
and underpowered.

Keywords
Atypical antipsychotics, ischemic stroke, case-crossover, national health insurance claims database

Introduction
Atypical antipsychotics (AAPs) have been widely used in the
treatment of schizophrenia, acute mania, bipolar disorder, behav-
ioral and psychological symptoms of dementia (BPSD) and delir-
ium since the 1990s (Horacek et al., 2006). They have been
effective for both positive and negative psychotic symptoms with-
out adverse events, including extrapyramidal symptoms, hypoten-
sion and anticholinergic-like syndrome (Devanand et al., 1998;
Small et al., 1997). Consensus guidelines, developed in 2004 to
guide treatment of psychosis associated with agitation or aggres-
sion in dementia (Alexopoulos et al., 2004), suggested that AAPs
were first-line choices due to their multiple sites of action and the
complexity of their changes responsible for neuropsychiatric
symptoms particularly in the treatment of BPSD in the elderly
(Mazzucco et al., 2008). Therefore, the use of AAPs has increased
markedly in the global prescription market (Kales et al., 2011).
Despite their increased use, the health regulatory agency of
guidelines emphasize the importance of generating evidence for
AAPs (Barnes, 2011; Burke and Tariot, 2009).
Of particular note, ischemic stroke has been reported to be
highly fatal among CVAEs and has been reported as a major cause
of death and disability worldwide (Donnan et al., 2008). A recent
systematic review showed conflicting findings regarding the cer-
ebrovascular safety of risperidone (Mazzucco et al., 2008;
Sacchetti et al., 2010b). Additionally, data from a clinical trial
using quetiapine to treat BPSD showed no increased risk of
1Office of Drug Utilization Review, Korea Institute of Drug Safety and
Risk Management, Seoul, Republic of Korea
2Medical Research Collaborating Center, Seoul National University
College of Medicine, Seoul, Republic of Korea
3Office of Pharmacoepidemiology, Korea Institute of Drug Safety and

Canada (2002) (Wooltorton, 2004) and the US Food and Drug Risk Management, Seoul, Republic of Korea
Administration (FDA) (2003) (McConville and Sorter, 2004) 4Department of Neuropsychiatry, Seoul National University Hospital,

reported an increased risk of cerebrovascular adverse events Seoul, Republic of Korea

5Department of Preventive Medicine, Seoul National University College
(CVAEs) from olanzapine and risperidone in clinical trials of
elderly demented patients, requiring label changes in prescribing of Medicine, Seoul, Republic of Korea
information. The European Medicines Agency (EMA) also Corresponding author:
reported that olanzapine use led to an increased risk of CVAEs and Byung-Joo Park, Department of Preventive Medicine, Seoul National
mortality in elderly people with dementia (Mowat et al., 2004). University College of Medicine, 103 Daehakno, Chongro-Gu, Seoul 110-
Since these regulatory agency warnings, there have been concerns 799, Republic of Korea.
about the risk of CVAEs associated with AAPs and recent Email: bjpark@snu.ac.kr

Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015

Shin et al.
CVAEs compared with a placebo (Tariot et al., 2006; Zhong et al.,
2007). Another clinical trial also suggested that olanzapine could
be a safe medication in the elderly (Moretti et al., 2005). However,
based upon limited evidence, further studies using a larger data-
base and addressing elderly patients should be conducted to con-
firm any such association.
In the current study, we used a case-crossover design to assess
the risk of ischemic stroke associated with risperidone, quetiapine
and olanzapine by eliminating confounders between individuals.
Using this technique, we estimated whether ischemic stroke had
been associated with the use of these drugs in a large elderly popu-
lation, and whether the risk of stroke was different among patients
with and without dementia.
Methods
Data source
The Korean Health Insurance Review and Assessment Service
(HIRA) database was used for this study. The National Health
Insurance (NHI) program was initiated in Korea in 1977 and
achieved universal coverage of the entire Korean population by
1989. Accordingly, the HIRA database contains all information on
the diagnoses and prescribed medications for approximately 50
million Koreans. We obtained claims data for elderly patients (aged
65 years or older) that had been submitted by healthcare providers
between 1 January 2005–30 June 2006. The database contained
information regarding 4,159,305 elderly patients with 100,838,744
prescriptions. The HIRA provided the claims data with the indi-
vidual identity of patients concealed. The database included an
anonymized code representing each individual together with their
age, gender, diagnoses and prescription drugs. Information on a
prescribed drug included the brand name, generic name, prescrip-
tion date, duration and route of administration. All diagnoses had
been coded according to the International Classification of Disease,
Tenth Revision (ICD-10). This study was approved by the
Institutional Review Board of Seoul National University College
of Medicine/Seoul National University Hospital.
Study subjects
The study population consisted of elderly patients who had
received atypical antipsychotics prior to their first ischemic-stroke
related hospitalization (ICD-10: I63) from 1 July 2005–30 June
2006. The index date was defined as the first hospital admission
date for ischemic stroke. We excluded the following patients: (a)
those with pre-existing cerebrovascular diseases (CVD) (ICD-10:
Figure 1. Definition of hazard and control periods in this case-cross study.
Hazard period was defined as the 30, 20, 10 day windows before the index date. A 30-day washout period was chosen between the end of the hazard period and the start
of each control period. Two control periods were defined as 30, 20, and 10 day windows to match those of the hazard period.
Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015
639
I60–I69) or transient ischemic attack (ICD-10: G45) during the
preceding six months from 1 January 2005–30 June 2005; (b)
those who had been diagnosed with ischemic stroke recorded the
same day as the first AAPs prescriptions; (c) those who were older
than 100 years; and (d) those who were receiving non-oral AAPs
(injectable or depot preparation).
Study design
We employed a case-crossover approach (Maclure, 1991) to
assess the association between AAP use and the risk of ischemic
stroke for controlling confounders within subjects. Instead of
matched controls, using study subjects as their own controls ena-
bled us to reduce confounding by subject characteristics that did
not change over time, including characteristics that were difficult
to investigate or impossible to measure (Delaney and Suissa,
2009). In this design, only patients experiencing an event of inter-
est were included and their exposures were measured during the
case and control periods.
The lengths of hazard periods were chosen based on the
criteria suggested by Maclure and Mittleman, i.e. the one that
maximized the odds ratio (Maclure and Mittleman, 2000). In
previous research related to AAPs and increased risk of stroke,
the highest risk was reported during the first weeks of treat-
ment (Kleijeret al., 2009) or within one month compared to the
period after one month (Sacchetti et al., 2010a). Therefore, we
investigated various definitions of the hazard period (i.e. 30
days, 20 days, 10 days) to determine the period over which
assessment of drug exposure occurred. All AAPs, including
risperidone, quetiapine and olanzapine were assessed within
30, 20, and 10 days prior to the incident stroke (hazard period),
respectively.
We set two control periods whose lengths were the same as the
hazard period. Two control periods were matched to enhance the
strength of the analyses as 60–90 and 120–150 days before the event.
A 30-day washout period was chosen between the end of the
case period and the start of the control period. Accordingly, for
each patient, the AAP prescription in each window period prior to
hospitalization for the ischemic stroke was compared with AAP
prescription in two earlier control-window periods. A washout
period was established between the case period and each of the two
control periods to reduce the likelihood of overlapping prescrip-
tions between these periods. The mean duration (± standard devia-
tion (SD)) of prescriptions for AAPs was 21.8±15.7 days and
prescriptions with a duration of less than 60 days consisted of
98.7% of total prescriptions. Therefore, we selected 30 days as the
washout period (Figure 1).
640 Journal of Psychopharmacology 27(7)

Statistical analyses users also showed the highest risk at 30 days (aOR=2.7, 95% CI:
2.0–3.6). Olanzapine did not show a statistically significant risk
Descriptive statistics including frequencies, proportions and for ischemic stroke (Table 2).
means (±SD) were used to illustrate the characteristics of the When the association was stratified by gender, risperidone
study subject by age, gender, and frequently prescribed AAPs. showed a higher risk in females (aOR=3.9, 95% CI: 3.1–5.0),
The distribution of psychiatric characteristics and comorbidity whereas quetiapine showed a higher risk in males (aOR=3.2, 95%
were analyzed by frequency and proportion. The presence of CI: 1.9–5.1). Risperidone and quetiapine showed the highest risk
dementia (F00–F03, G30, G31.8), psychiatric characteristics and for ischemic stroke in both demented and non-demented patients.
comorbidity information was extracted by the ICD-10 codes On the other hand, olanzapine showed a significant increased risk
recorded during the six months prior to the index date for each only in demented patients (Table 3).
case. Psychiatric characteristics included depression (F32, F33,
F34.1, F41.2), schizophrenia (F20), bipolar affective disorder
(F31), manic episode (F30), other psychotic disorder (F22.0, F23, Discussion
F29), obsessive-compulsive disorder (F42) and delirium (F05).
Comorbidity included essential hypertension (I10–15), diabetes This population-based case-crossover study revealed that risperi-
mellitus (E10–E14), chronic obstructive pulmonary disease done and quetiapine use in the most recent 30 days was associated
(COPD) (J40–44, J47), coronary heart disease (I20–25), dyslipi- with increased risk of ischemic stroke in patients. Risperidone and
demia (E78) and heart failure (I50). Prescriptions of any AAPs, quetiapine showed similar increased risk profiles in geriatric
risperidone, quetiapine and olanzapine between case and control patients regardless of the presence of dementia. On the other hand,
periods were compared and statistical significance was examined olanzapine was significantly associated with risk of ischemic
using McNemar’s test. Odds ratios (ORs) for stroke associated stroke only in demented geriatric patients.
with the use of AAPs and 95% confidence intervals (CIs) were
calculated by conditional logistic regression.
In each patient, if acute comorbidity or comedication that Table 1. Characteristics of elderly patients with prescriptions of
could increase stroke risk was claimed to vary between the hazard atypical antipsychotics prior to hospitalization for ischemic stroke.
and control periods, we adjusted the factors in the regression
model. Potential confounders included a range of variables (see Characteristics No. of subjects %
Appendix 1). We assessed AAP prescriptions, the diagnosis of (n=1601)
comorbidity and the use of comedication included in the hazard or Mean age (yr) (±SD) 75.6 (±6.7)
control periods. Exposure to AAPs, comorbidity and comedica-
65–74 1313 46.4
tion were considered to be dichotomous variables in the model
75–84 1239 43.8
(exposed at least once during each specific time window: yes or
Over 85 277 9.8
no). We included variables achieving statistical significance in the
Gender
likelihood ratio test in the final model. Finally, the model included
Male 668 41.7
antidepressants, benzodiazepine, anticoagulants, and the diagno-
sis of pneumonia, acute myocardial infarction and atrial fibrilla- Female 933 58.3
tion as the adjusting variables. Stratified analysis was also Atypical antipsychotics
performed to assess the effect modification according to age Risperidone 1154 64.7
group, gender, the presence or absence of dementia diagnosis, Quetiapine 429 24.1
psychiatric characteristics (depression, schizophrenia, and bipolar Olanzapine 146 8.2
affective disorder) and chronic disease status (essential hyperten- Others 54 3.0
sion, diabetes mellitus, COPD, coronary heart disease, dyslipi- Diagnosis of dementia
demia, and heart failure). Statistical analysis was performed using Yes 954 59.6
the SAS statistical application program (Release 9.2, SAS Institute No 647 40.4
Inc., Cary, North Carolina, USA). A two-tailed value of p<0.05 Psychiatric characteristics
was considered statistically significant. Depression 512 32.0
Schizophrenia 315 19.7
Bipolar affective disorder 128 8.0
Results Manic episode 57 3.6
We identified 454,049 patients aged older than 64 years who had Other psychotic disorder 617 38.5
been hospitalized for ischemic stroke from 1 July 2005–30 June Obsessive-compulsive disorder 17 1.1
2006. A total of 1601 patients with at least one prescription for an Delirium 389 24.3
AAP prior to incident ischemic stroke cases were included in the Chronic comorbidity
final analysis (Figure 2). A total of 954 patients (59.6%) had Essential hypertension 1015 63.4
received a diagnosis of dementia in the previous six months before Diabetes mellitus 484 30.2
the incident ischemic stroke (Table 1). The adjusted ORs for COPD 286 17.9
ischemic stroke associated with any AAP were 3.9 (95% CI: 3.3– Coronary heart disease 278 17.4
4.6), 3.6 (95% CI: 3.0–4.3), and 2.8 (95% CI: 2.3–3.3), respec- Dyslipidemia 235 14.7
tively. In risperidone users, the highest risk was observed in the 30
days of risk window (aOR=3.5, 95% CI: 2.9–4.2), and quetiapine COPD: chronic obstructive pulmonary disease; SD: standard deviation.

Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015

Shin et al. 641

Table 2. Risk of ischemic stroke related to the use of risperidone, quetiapine and olanzapine in the elderly.

Time window period No. exposed in hazard No. exposed in control Unadjusted Adjusted ORb
period (n=1601) period (n=3202) ORa (95% CI) (95% CI)

Exposure assessed within 30 days

Any AAP user 998 1119 3.9 (3.4–4.5) 3.9 (3.3–4.6)
Risperidone 687 717 3.9 (3.3–4.7) 3.5 (2.9–4.2)
Quetiapine 250 292 3.5 (2.6–4.6) 2.7 (2.0–3.6)
Olanzapine 77 112 2.1 (1.4–3.4) 1.2 (0.7–2.0)
Exposure assessed within 20 days
Any AAP user 924 1059 3.7 (3.2–4.3) 3.6 (3.0–4.3)
Risperidone 636 676 3.7 (3.1–4.4) 3.4 (2.8–4.0)
Quetiapine 226 270 3.3 (2.5–4.3) 2.6 (2.0–3.6)
Olanzapine 70 107 1.9 (1.2–3.1) 1.3 (0.8–2.1)
Exposure assessed within 10 days
Any AAP user 799 976 3.0 (2.6–3.5) 2.8 (2.3–3.3)
Risperidone 538 614 3.0 (2.6–3.6) 2.7 (2.2–3.2)
Quetiapine 201 256 2.7 (2.1–3.7) 2.3 (1.7–3.1)
Olanzapine 60 96 1.8 (1.1–3.1) 1.2 (0.7–2.1)

AAP: atypical antipsychotic; CI: confidence interval; OR: odds ratio.

aCalculated by conditional logistic regression.
bCalculated by conditional logistic regression adjusted for the discordant diagnosis of acute myocardial infarction, pneumonia, and atrial fibrillation

and the use of antidepressants, benzodiazepine, and anticoagulants.

Figure 2. The selection of study subjects. AAP: atypical antipsychotic; ICD-10: International Classification of Disease, Tenth Revision.

Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015

642 Journal of Psychopharmacology 27(7)

Table 3. Risk of ischemic stroke associated with the use of risperidone, quetiapine, and olanzapine in the previous 30 days, total risk and stratified
analysis according to age group, gender, the presence of dementia, psychiatric characteristics and comorbidity.

No. of cases Any AAP users Risperidone Quetiapine (n=429) Olanzapine

(n=1601) (n=1154) (n=146)

Total 1601 3.9 (3.3–4.6) 3.5(2.9–4.2) 2.7 (2.0–3.6) 1.2 (0.7–2)

Age, years
65–74 1313 4.3 (3.3–5.5) 3.9 (2.9–5.2) 3.3 (2.1–5.0) a

75–84 1239 3.3 (2.6–4.3) 3.1 (2.4–4.1) 2.0 (1.3–3.1) 2.1 (1.0–4.7)
85+ 277 5.0 (2.9–8.6) 3.5 (2.0–6.1) 4.8 (1.8–12.9) a

Gender
Male 668 3.3 (2.6–4.3) 2.9 (2.2–3.9) 3.2 (1.9–5.1) 1.1 (0.5–2.1)
Female 933 4.4 (3.5–5.5) 3.9 (3.1–5.0) 2.5 (1.7–3.6) 1.9 (0.9–3.8)
Dementia
Yes 954 3.9 (3.2–4.9) 3.6 (2.9–4.5) 2.5 (1.7–3.6) 3.0 (1.1–8.1)
No 647 3.8 (2.9–5.0) 3.4 (2.4–4.6) 3.0 (1.9–4.8) 1.0 (0.6–1.9)
Comorbidity
Depression 512 3.4 (2.5–4.5) 3.6 (2.6–5.1) 2.1 (1.3–3.3) 0.8 (0.3–1.8)
Schizophrenia 315 3.3 (2.2–4.8) 3.1 (2.0–4.9) 2.1 (1.3–3.5) 1.5 (0.7–6.4)
BAD 128 4.4 (2.4–8.0) 4.1 (1.9–8.5) 5.3 (2.0–13.7) 3.0 (1.1–8.1)
Hypertension 1015 4.0 (3.3–5.0) 3.1 (2.5–3.9) 3.2 (2.2–4.7) 2.2 (1.1–4.2)
DM 484 4.4 (3.2–6.1) 3.3 (2.4–4.7) 4.4 (2.6–7.6) 1.1 (0.5–2.6)
COPD 286 4.3 (2.8–6.4) 3.8 (2.5–5.9) 3.3 (1.4–7.8) 1.4 (0.5–3.7)
CHD 278 4.3 (2.8–6.6 3.1 (2.0–4.7) 3.1 (1.7–5.8) 3.1 (0.7–13.1)
Dyslipidemia 235 7.2 (4.4–11.8) 7.1 (4.1–12.4) 3.9 (1.9–8.0) 1.6 (0.2–12.4)

AAP: atypical antipsychotic; BAD: bipolar affective disorder; CHD: coronary heart disease; COPD: chronic obstructive pulmonary disease; DM: diabetes mellitus.
aCould not be estimated.

Our results showed nearly a four-fold increase in the risk of
ischemic stroke related to the use of all AAPs. Previous self-
controlled case series study also showed a 2.3-fold increase in risk
with AAPs (Douglas and Smeeth, 2008), and a previous retro-
spective cohort study showed a 2.5-fold increase in risk based on
AAP use compared to the no use group, the 95% CIs for these
estimates fully overlapped with our results (Sacchetti et al., 2008).
The estimated risk in risperidone users was generally consistent
with the results of previous studies, supporting a previous meta-
analysis which showed a similar 3.4-fold estimated risk (Schneider
et al., 2006a) and randomized controlled trials (RCTs) (Brodaty
et al., 2005; Deberdt et al., 2005). Quetiapine use also showed a
2.7-fold increased risk of ischemic stroke, supporting previous
cohort study which showed a 2.1-fold increased risk compared to
olanzapine; however, one RCT demonstrated the safety of quetia-
pine by showing similar rates of CVAEs between placebo and
quetiapine users (Zhong et al., 2007).
Olanzapine did not increase the risk of ischemic stroke in geri-
atric patients: however, the number of olanzapine-exposed
patients in the hazard period or control period was 91 in the
demented (power=51.2%), and 63 in the non-demented patients
(power=5.0%), respectively. The study was underpowered due to
the small sample size. Olanzapine was significantly associated
with a risk of ischemic stroke only in demented patients when
performing stratified-analysis. Previous RCTs have also shown
increased CVAE event rates for olanzapine compared to risperi-
done (Deberdt et al., 2005) or quetiapine (Schneider et al., 2006b)
in patients with dementia. However, one controlled open-label
study which compared olanzapine with typical antipsychotics
reported that olanzapine could be a safe treatment (Moretti et al.,
2005). Cohort studies have also shown olanzapine to have a 1.1-
fold increased risk compared to typical antipsychotics (Herrmann
et al., 2004), and a 0.8-fold decreased risk compared to risperi-
done (Finkel et al., 2005). However, the findings in these two pre-
vious studies were not statistically significant.
Although previous studies have conflicted, our results suggest
that olanzapine does not appear to be associated with ischemic
stroke relative to risperidone or quetiapine in non-demented
patients. However, it might be a higher risk in geriatric patients
with dementia. Our study showed that both demented and non-
demented patients had similar risk profiles for ischemic stroke in
risperidone and quetiapine users. This finding was consistent with
previous research which showed similar hazard ratios for the risk
of death regardless of having dementia or not (Wang et al., 2005).
However another study showed an increased risk in demented
patients (Douglas and Smeeth, 2008) and most previous research
included the study subjects as the patients with dementia. Our
results suggest that prescribing AAPs, particularly risperidone or
quetiapine, should be done with more caution in non-demented
geriatric patients. More data is needed to confirm the risk in geri-
atric patients without dementia.
Biological mechanism
Up to now, the biological mechanisms responsible for a possi-
ble increased risk of ischemic stroke have remained unknown
although several hypotheses have been suggested (Wada-Isoe
et al., 2004). Our study results imply that a transient effect or an
antipsychotic effect on glucose or lipid metabolism may not be
involved as possible mechanisms. Even though risperidone,

Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015

Shin et al.
quetiapine and olanzapine have all been categorized as AAPs,
we found substantially different results: differences that may
relate to their chemical class, structure and mechanism.
Particularly, differences in safety profiles related to hyperprol-
actinemia and hypotension among risperidone, quetiapine and
olanzapine have been demonstrated (McConville and Sorter,
2004). Hyperprolactinemia has been mostly seen with risperi-
done use, although olanzapine and quetiapine are also associ-
ated with hyperprolactinemia and related symptoms, albeit to a
lesser extent. This implies that drug-induced hyperprolactine-
mia, which may promote platelet aggregation (Smith and Beier,
2004; Wallaschofski et al., 2001), could increase the risk of
ischemic stroke.
Gender differences in the stratified analysis between risperi-
done and quetiapine might be related with the hyperprolactinemia
hypothesis. Normally, females are more sensitive to the risk of
hyperprolactinemia (Kinon et al., 2003), which may explain our
results showing that female risperidone users were the most at
risk, whereas quetiapine was associated with a higher risk in
males. Orthostatic hypotension also has been mostly seen with
risperidone use and may explain the fact that direct cardiovascular
effects are related to increased risk of ischemic stroke (Gardner
et al., 2005).
This study had several strengths. Our study included the entire
South Korean population rather than a sample and used the claims
database for all incident cases of ischemic stroke collected by the
NHI program, which covered nearly all inhabitants of Korea. In
addition, we effectively controlled for unmeasured confounders
which were stable over time, applying a case-crossover design.
Case-crossover designs have been deemed particularly suitable
when the exposure’s effect on risk is immediate and transient, and
the outcome is abrupt (Maclure, 1991). In the present study, using
the case-crossover technique, only cases with incident stroke were
considered, and their AAP exposures were compared during three
different time windows. Since inherent confounders remain invar-
iant over time, the case-crossover design has been shown to mini-
mize between-subject confounding and assure an optimal sample
size (Delaney and Suissa, 2009). In the current study, we further
adjusted for other medication use that could be associated with the
development of ischemic stroke. We also provided more detailed
information about the timing of stroke and risk profiles in patients
with and without dementia.
Despite our findings, our results should be interpreted with
caution. The incident cases of ischemic stroke were defined
according to the diagnosis reported to the HIRA, which might
have differed from the patients’ actual diagnoses. A previous vali-
dation study compared the diagnosis derived from the HIRA data-
base with the actual diagnosis in patient medical records. The
overall positive predictive value of the diagnoses was 83.4% in
cases of hospitalized patients. Any misclassification of exposure
status that occurred non-differentially would have produced
biased results toward the null: differential misclassification such
as differences in adherence between AAPs would be similar
between the case and control period, which would affect the esti-
mated ratio toward the null. Also, we could not identify the time-
variant clinical characteristics that might have been different
between the case and control periods, given that the HIRA data-
base did not include this information. For a more in-depth investi-
gation, laboratory data within each hospital would have to be
linked with the HIRA database.
Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015
643
In conclusion, we found a provocative tripled risk of ischemic
stroke following risperidone and quetiapine use: however, an
increased stroke risk in demented patients, assessed within 30
days after exposure, was also observed with olanzapine. Both
demented and non-demented patients showed a similar increased
risk profile with risperidone and quetiapine use. Based on this,
prescribing risperidone and quetiapine should be done with
increased caution regardless of the presence or absence of demen-
tia. Further, olanzapine may be more acceptable in non-demented
geriatric patients associated with the risk of ischemic stroke,
although wider consideration of all potential risks and benefits
should be considered.
Acknowledgements
The authors thank HIRA of Korea and their related persons for acquisition
of the data.
Conflict of interest
The authors declare that there are no conflicts of interest.
Funding
This research received no specific grant from any funding agency in the
public, commercial, or not-for-profit sectors.
References
Alexopoulos GS, Streim J, Carpenter D, et al.; Expert Consensus Panel
for Using Antipsychotic Drugs in Older Patients. Using antipsychotic
agents in older patients. J Clin Psychiatry 2004; 65: S5–S99.
Barnes RET; Schizophrenia Consensus Group of the British Association
for Psychopharmacology. (2011) Evidence-based guidelines for the
pharmacological treatment of schizophrenia: Recommendations from
the British Association for Psychopharmacology. J Psychopharmacol
25: 567–620.
Brodaty H, Ames D, Snowdon J, et al. (2005) Risperidone for
psychosis of Alzheimer's disease and mixed dementia: Results of
a double-blind, placebo-controlled trial. Int J Geriatr Psychiatry
20: 1153–1157.
Burke AD and Tariot PD (2009) Atypical antipsychotics in the elderly:
A review of therapeutic trends and clinical outcomes. Expert Opin
Pharmacother 10: 2407–2414.
Deberdt WG, Dysken MW, Rappaport SA, et al. (2005) Comparison of
olanzapine and risperidone in the treatment of psychosis and associ-
ated behavioral disturbances in patients with dementia. Am J Geriatr
Psychiatry 13: 722–730.
Delaney JA and Suissa S (2009) The case-crossover study design in
pharmacoepidemiology. Stat Methods Med Res 18: 53–65.
Devanand DP, Marder K, Michaels KS, et al. (1998) A randomized,
placebo-controlled dose-comparison trial of haloperidol for psycho-
sis and disruptive behaviors in Alzheimer's disease. Am J Psychiatry
155: 1512–1520.
Donnan GA, Fisher M, Macleod M, et al. (2008) Stroke. Lancet 371:
1612–1623.
Douglas IJ and Smeeth L (2008) Exposure to antipsychotics and risk of
stroke: Self controlled case series study. Brit Med J 337: a1227.
Finkel S, Kozma C, Long S, et al. (2005) Risperidone treatment in elderly
patients with dementia: Relative risk of cerebrovascular events versus
other antipsychotics. Int Psychogeriatr 17: 617–629.
Gardner DM, Baldessarini RJ and Waraich P (2005) Modern antipsy-
chotic drugs: A critical overview. CMAJ 172: 1703–1711.
Herrmann N, Mamdani M and Lanctot KL (2004) Atypical antipsychot-
ics and risk of cerebrovascular accidents. Am J Psychiatry 161:
1113–1115.
644 Journal of Psychopharmacology 27(7)

Horacek J, Bubenikova-Valesova V, Kopecek M, et al. (2006) Mecha-
nism of action of atypical antipsychotic drugs and the neurobiology
of schizophrenia. CNS Drugs 20: 389–409.
Kales HC, Zivin K, Kim HM, et al. (2011) Trends in antipsychotic use in
dementia 1999–2007. Arch Gen Psychiatry 68: 190–197.
Kinon BJ, Gilmore JA, Liu H, et al. (2003) Hyperprolactinemia in
response to antipsychotic drugs: Characterization across comparative
clinical trials. Psychoneuroendocrinology 28: S69–S82.
Kleijer BC, van Marum RJ, Egberts ACG, et al. (2009) Risk of cerebro-
vascular events in elderly users of antipsychotics. J Psychopharmacol
23: 909–914.
McConville B and Sorter M (2004) Treatment challenges and safety
considerations for antipsychotic use in children and adolescents with
psychoses. J Clin Psychiatry 65: S20–S29.
Maclure M (1991) The case-crossover design: A method for studying
transient effects on the risk of acute events. Am J Epidemiol 133:
144–153.
Maclure M and Mittleman MA (2000) Should we use a case-crossover
design? Annu Rev Public Health 21: 193–221.
Mazzucco S, Cipriani A, Barbui C, et al. (2008) Antipsychotic drugs and
cerebrovascular events in elderly patients with dementia: A system-
atic review. Mini Rev Med Chem 8: 776–783.
Moretti R, Torre P, Antonello R, et al. (2005) Olanzapine as a possible
treatment of behavioral symptoms in vascular dementia: Risks of
cerebrovascular events. J Neurol 252: 1186–1193.
Mowat D, Fowlie D and Macewan T (2004) Csm warning on atypical
psychotics and stroke may be detrimental for dementia. Brit Med J
328: 1262.
Sacchetti E, Trifiro G, Caputi A, et al. (2008) Risk of stroke with typical
and atypical anti-psychotics: A retrospective cohort study including
unexposed subjects. J Psychopharmacol 22: 39–46.
Sacchetti E, Turrina C, Cesana B, et al. (2010a) Timing of stroke in elderly
people exposed to typical and atypical antipsychotics: A replication
cohort study after the paper of Kleijer, et al. J Psychopharmacol 24:
1131–1132.
Sacchetti E, Turrina C and Valsecchi P (2010b) Cerebrovascular acci-
dents in elderly people treated with antipsychotic drugs: A systematic
review. Drug Saf 33: 273–288.
Schneider LS, Dagerman K and Insel PS (2006a) Efficacy and adverse
effects of atypical antipsychotics for dementia: Meta-analysis of
randomized, placebo-controlled trials. Am J Geriatr Psychiatry 14:
191–210.
Schneider LS, Tariot PN, Dagerman KS, et al. (2006b) Effectiveness of
atypical antipsychotic drugs in patients with Alzheimer's disease. N
Engl J Med 355: 1525–1538.
Small GW, Rabins PV, Barry PP, et al. (1997) Diagnosis and treatment of
Alzheimer disease and related disorders. Consensus statement of the
American Association for Geriatric Psychiatry, the Alzheimer's Asso-
ciation, and the American Geriatrics Society. JAMA 278: 1363–1371.
Smith DA and Beier MT (2004) Association between risperidone treat-
ment and cerebrovascular adverse events: Examining the evidence
and postulating hypotheses for an underlying mechanism. J Am Med
Dir Assoc 5: 129–132.
Tariot PN, Schneider L, Katz IR, et al. (2006) Quetiapine treatment of psy-
chosis associated with dementia: A double-blind, randomized, placebo-
controlled clinical trial. Am J Geriatr Psychiatry 14: 767–776.
Wada-Isoe K, Wakutani Y, Urakami K, et al. (2004) Elevated interleu-
kin-6 levels in cerebrospinal fluid of vascular dementia patients. Acta
Neurol Scand 110: 124–127.
Wallaschofski H, Donne M, Eigenthaler M, et al. (2001) PRL as a novel
potent cofactor for platelet aggregation. J Clin Endocrinol Metab 86:
5912–5919.
Wang PS, Schneeweiss S, Avorn J, et al. (2005) Risk of death in elderly
users of conventional vs. atypical antipsychotic medications. N Engl
J Med 353: 2335–2341.
Wooltorton E (2004) Olanzapine (Zyprexa): Increased incidence of cere-
brovascular events in dementia trials. CMAJ 170: 1395.
Zhong KX, Tariot PN, Mintzer J, et al. (2007) Quetiapine to treat agitation
in dementia: A randomized, double-blind, placebo-controlled study.
Curr Alzheimer Res 4: 81–93.

Appendix 1
Potential confounders including the discordant use of comedication and comorbidity.

Discordant use of comedication

Benzodiazepines alprazolam, bromazepam, brotizolam, clotiazepam, etizolam, fludiazepam, midazolam, oxazepam,
tofisopam, clobazam, clonazepam, clorazepate dipotassium, chlordiazepoxide, diazepam, estazolam, ethyl
loflazepate, flunitrazepam, flurazepam, mexazolam, nordazepam, pinazepam
Antidepressants amitriptyline, amoxapine, bupropion, citalopram, clomipramine, dothiepin, escitalopram, fluoxetine,
fluvoxamine, imipramine, mianserin, milnacipran, mirtazapine, nortryptyline, paroxetine, quinupramine,
sertraline, trazodone, venlafaxine
Anticoagulants warfarin, heparin, antithrombin III, dalteparin, enoxaparin, nadroparin, parnaparin, bemiparin
Antiplatelets clopidogrel, dipyridamole, ticlopidine, aspirin, indobufen, abciximab, tirofiban, triflusal, ozagrel,
sarpogrelate, cilostazol
Other antithrombotic agents streptokinase, alteplase, urokinase, lepirudin, argatroban, rivaroxaban
Diagnosis of comorbidity
Manic episode ICD-10 Code F30
Other psychotic disorder ICD-10 Code F22.0, F23, F29
Obsessive-compulsive disorder ICD-10 Code F42
Delirium ICD-10 Code F05
Infective endocarditis ICD-10 Code I33.0
Valvular heart disease ICD-10 Code I06
Acute myocardial infarction ICD-10 Code I21
Thyrotoxicosis ICD-10 Code E05
Pneumonia ICD-10 Code J12
Vulvar thromboembolism ICD-10 Code I24.0
Atrial fibrillation ICD-10 Code I45

Downloaded from jop.sagepub.com at Yale University Library on February 4, 2015