JID: NEUPSY

ARTICLE IN PRESS [m6+;September 19, 2018;17:35]

European Neuropsychopharmacology (2018) 000, 1–11

www.elsevier.com/locate/euroneuro

Antipsychotic drugs for elderly patients

with schizophrenia: A systematic review
and meta-analysis
Marc Krause a,c,∗, Maximilian Huhn a,
Johannes Schneider-Thoma a, Philipp Rothe a, Robert C Smith b,
Stefan Leucht a

a
Department of Psychiatry and Psychotherapy, Technical University of Munich, Klinikum rechts der Isar,
Ismaningerstraße 22, 81675 Munich, Germany
b
New York University Medical School, Department Of Psychiatry, and Nathan Kline Institute for
Psychiatric Research, New York, USA
c
Ludwig-Maximilians Universität München, Germany

Received 17 May 2018; accepted 5 September 2018

Available online xxx

KEYWORDS Abstract
Elderly; Elderly patients with schizophrenia are a particularly vulnerable group often excluded from
Schizophrenia; clinical trials. Currently there is no evidence-synthesis about the efficacy and safety of antipsy-
Antipsychotics; chotics in this subgroup.
Meta-analysis We reviewed all randomized-controlled-trials, about antipsychotics in elderly schizophrenics
(last search Dec 12, 2017). Pairwise meta-analyses were conducted. The primary outcome was
overall symptoms. Secondary outcomes included positive symptoms, negative symptoms, re-
sponse, dropouts, quality of life, social functioning and side-effects.
We included 29 references from 18 unique randomized-controlled-trials with 1225 participants
published from 1958 to 2009. The definition of “elderly” was very heterogeneous across the
studies (minimum age 46–65, mean age 57–73). There were evidence gaps for most drugs in
many outcomes. In terms of efficacy paliperidone was associated with fewer dropouts due
to inefficacy than placebo in the only placebo-controlled-trial. Olanzapine was superior to
haloperidol in overall symptoms, negative symptoms and response, and it was associated with
fewer dropouts than risperidone. Risperidone and haloperidol produced more prolactin increase
than olanzapine, and olanzapine was associated with less use of antiparkinson medication than
haloperidol.

∗ Corresponding author at: Department of Psychiatry and Psychotherapy, Technical University of Munich. Klinikum rechts der Isar, Is-

maningerstraße 22, 81675 Munich, Germany.

E-mail address: marc.krause@tum.de (M. Krause).
https://doi.org/10.1016/j.euroneuro.2018.09.007
0924-977X/© 2018 Elsevier B.V. and ECNP. All rights reserved.

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS [m6+;September 19, 2018;17:35]
2 M. Krause et al.

Although we found no marked differences of the effects of these drugs in the elderly, the evi-
dence presented was based on very few usually small studies. To examine specifically whether
there are differences in efficacy and side-effects in elderly, which differs in meaningful ways
from the general population, studies in patients who are defined by critiera as truly geriatric,
which incorporates older age together with multimorbidity and fraility dimensions, may be
more informative.
© 2018 Elsevier B.V. and ECNP. All rights reserved.

1. Introduction ficacious than placebo, and whether there are differences

Aging is a global trend, according to studies by the United
Nations Population Department (United Nations, Depart-
ment of Economic and Social Affairs, Population Division,
2015), and demographic trends show an increase in the old
and very old in many countries.
The life morbidity risk for schizophrenia is approximately
1% (mean/median 11/7 per 1000), the median point preva-
lence per 1000 is 4.6 (10, 90 percent quantiles 1.9, 10.0),
and the median yearly incidence per 100.000 is 15.2 (10,
90 percent quantiles 7.7, 43.0) (McGrath et al., 2008).
Schizophrenia first appears in adolescence but continues
throughout the life cycle. The life-time prevalence of
schizophrenia in old age fluctuates between 0.5 and 1%
(Andreas et al., 2017). Treatment with antipsychotic drugs is
the primary therapy for patients with schizophrenia (Leucht
et al., 2012) and they are widely used in older patients
(Colenda et al., 2002). However, the efficacy and side-
effects of antipsychotic drugs for treatment of schizophre-
nia in older patients has not been systematically evaluated.
Side-effects of antipsychotic drugs which can include weight
gain, glucose-lipid abnormalities and hormonal changes can
be especially important because they can increase cardio-
vascular and other risk factors which are more important in
the elderly population.
A systematic evaluation is important to provide a data-
base to facilitate evidence-based decisions for selecting
the best individual treatment for elderly patients with
schizophrenia. However most of the clinical studies have ex-
cluded these patients, because their inclusion criteria were
often restricted to patients up to an age of sixty-five years.
This also applies to systematic reviews and meta-analyses,
which deliberately exclude the few available studies that
examine the subgroup of the old patients with schizophre-
nia (Leucht et al., 2013). Although older schizophrenics
are often treated with antipsychotics, the available evi-
dence is insufficient to provide data on which evidence-
based clinical recommendations can be made. There is only
one Cochrane Review about the antipsychotic treatment of
patients with late onset schizophrenia, but nothing about
elderly patients with schizophrenia in general, independent
of the age of onset (Essali and Ali, 2012). However, the el-
derly may be more sensitive to antipsychotic drug side ef-
fects and may be less responsive to antipsychotic treatment
in some conditions. This background provides a rationale
for doing a separate meta-analysis of the efficacy and side-
effects of antipsychotics in the elderly. We, therefore, con-
ducted a systematic review and meta-analysis on the effects
of antipsychotics in elderly patients with schizophrenia. The
aims were to determine which antipsychotics are more ef-
between specific antipsychotics in efficacy and side-effects.
2. Methods
An a priori written study protocol was registered at PROS-
PERO under the registration number: CRD42016052060.
2.1. Participants and interventions
We included all randomized controlled trials (RCTs) in el-
derly patients with schizophrenia, schizophreniform disor-
der, or schizoaffective disorder (as defined by any diagnostic
criteria).
The interventions were 34 antipsychotic drugs which
comprised all second-generation antipsychotics available in
the US and/or Europe, and a selection of first-generation
antipsychotics, licensed in at least one country, which
were considered important based on a survey of interna-
tional schizophrenia experts (Leucht et al., 2016) (amisul-
pride, aripiprazole, asenapine, benperidol, brexpipra-
zole, cariprazine, chlorpromazine, clopenthixol, clozap-
ine, flupenthixol, fluphenazine, fluspirilene, haloperidol,
iloperidone, levomepromazine, loxapine, lurasidone, molin-
done, olanzapine, paliperidone, penfluridol, perazine, per-
phenazine, pimozide, quetiapine, risperidone, sertindole,
sulpiride, thioridazine, thiothixene, trifluoperazine, ziprasi-
done, zotepine, zuclopenthixol). We included these drugs
at any dose and in any form of administration when com-
pared with another antipsychotic or placebo, and used as
monotherapy. Drugs that failed to obtain FDA/EMA approval
or which had not yet been licensed were not considered.
2.2. Search strategy and selection criteria
We conducted a comprehensive, systematic literature
search in MEDLINE, EMBASE, PsycINFO, Cochrane Library,
PubMed, Biosis, and ClinicalTrials.gov up to Nov 17, 2016
(eAppendix 2 in the Supplement) and a final PubMed search
until Dec 12, 2017. The search terms included the generic
names of 34 antipsychotics listed above. The minimum dura-
tion of the RCTs was set at three weeks. We also inspected
the reference lists of the included studies and of previous
systematic reviews (Fusar-Poli et al., 2015; Leucht et al.,
2002) and a narrative review (Remington et al., 2016). Ci-
tations were screened independently by at least two re-
viewers (MK, MH, YZ) at both the title/abstract and full-
text stages. In the case of crossover studies, we only used

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS
Antipsychotic drugs for elderly patients with schizophrenia
the first crossover phase to avoid the problem of carry-
over effects (Elbourne et al., 2002). We excluded cluster-
randomized trials (Divine et al., 1992). Studies that demon-
strated a high risk of bias for sequence generation or allo-
cation concealment were excluded (Higgins, 2011). If a trial
was described as double-blind but randomization was not
explicitly mentioned, we assumed that study participants
were randomized, but we excluded the trial in a sensitivity
analysis. We excluded studies from mainland China to avoid
a systematic bias because many of these studies do not use
appropriate randomization procedures, do not report their
methods, and have been reported to often be not reliable
(Bian et al., 2006; Wu et al., 2009).
2.3. Outcome measures and data extraction
The primary outcome was the mean change from baseline to
endpoint in overall symptoms of schizophrenia as measured
by the Positive and Negative Syndrome Scale (PANSS) (Kay
et al., 1987), the Brief Psychiatric Rating Scale (Overall,
1962), or any other validated scale for the overall assess-
ment of schizophrenia symptoms, especially adapted ver-
sions of BPRS and PANSS for elderly. If change data were not
available, we used the mean score at study end point of
these scales. Intention-to-treat data sets were used when-
ever available.
Secondary outcomes were clinically important responses
to treatment defined by the authors, the mean change in
positive, negative and depressive symptoms of schizophre-
nia, dropouts owing to any reason (all-cause discontinua-
tion), dropouts owing to inefficacy of treatment and due
to adverse events, the occurrence of important adverse ef-
fects (weight gain, at least one extrapyramidal symptom,
sedation, akathisia, prolactin prolongation), antiparkinso-
nian medication used at least once as a proxy for extrapyra-
midal symptoms, quality of life, and social functioning.
We used the Cochrane Collaboration’s risk-of-bias tool for
the assessment of potential bias in terms of randomization,
allocation concealment, blinding, missing outcomes, selec-
tive reporting and other biases (Higgins, 2011). Data extrac-
tion and assessment of study quality were performed inde-
pendently by at least two reviewers (MK, MH, JS, PR, HR,
SB, LB, MS, TA, LR and NP). Disagreement was resolved by
discussion. If disagreement could not be resolved, we dis-
cussed with the team leader (SL). We sent emails to the first
and corresponding authors of all included studies to request
missing data. Missing SDs were estimated from test statistics
or substituted by the mean SD of the other included studies
using the same scale (Furukawa et al., 2006)
2.4. Statistical analysis
We originally planned to conduct a network meta-analysis.
However, the network plot was very poorly connected as
there were only a few eligible studies (see Fig. 1). Also,
there were neither triangular nor quadratic loops of direct
comparisons. Therefore, we decided to calculate pairwise,
random-effects meta-analyses using Review Manager 5.3.
This decision had been foreseen in our protocol. For con-
tinuous outcomes, the effect sizes were calculated as stan-
Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
[m6+;September 19, 2018;17:35]
3
dardized mean differences (SMDs) according to Hedges’s
g (Cohen, 2013). For binary outcomes, the effect sizes
were calculated as odds ratios (ORs) according to Mantel–
Haenszel. Both types of effect sizes were presented along
with their 95% confidence intervals (CIs). We applied the
random-effects model by DerSimonian and Laird (1986) to
all outcomes. Heterogeneity was assessed with the I² statis-
tic (Higgins, 2011) and a chi²-square test for homogeneity.
Small trial effects were explored by funnel-plots if at least
ten studies were available (Higgins, 2011). As the method
is based on symmetry, funnel-plots based on fewer trials
are not meaningful (Higgins, 2011). P-values lower than 0.05
were considered to be statistically significant.
3. Results
3.1. Description of included studies
We identified 29 references from 18 unique RCTs with
1225 unique participants published from 1958 to 2009. The
PRISMA flowchart is shown in Fig. 2. Details of all included
studies are presented in Table 1. Of 718 patients for whom
gender was indicated, 445 were women (62%). The mean
age of participants was 66.09 years (range 56.9–72.2 years).
The median trial duration was 10 weeks (range 3–72 weeks).
The assessment for risk of bias is presented in eAppendix
4 in the Supplement. Nearly all included studies reported
insufficient information concerning random sequence gen-
eration (17 of 18 studies) and allocation concealment (17 of
18 studies). The blinding of patients and personnel showed
a high risk of bias in five studies and low risk in 3 studies.
The risk of bias for blinding of outcome assessment was sim-
ilar with 4 studies at low risk and 3 at high risk. The number
of studies with high risk of bias for missing outcomes and
selective reporting were 5 and 6 respectively, and in both
categories six studies showed low risk of bias.
Fig. 1 shows the network of eligible comparisons for
the primary outcome. The identified studies included the
drugs amisulpride, carphenazine, chlorpromazine, clozap-
ine, fluphenazine, haloperidol, olanzapine, paliperidone,
placebo, quetiapine, risperidone, thioridazine, trifluoper-
azine. Two studies reported combined treatment groups
(typical neuroleptic, conventional). The drug involved in
most comparisons was risperidone (9 of 18 trials) fol-
lowed by haloperidol (7 of 18 trials) olanzapine (6 of 18
trials) and clozapine (3 of 18 trials), whereas just sin-
gle trials were available for other drugs: amisulpride,
carphenazine, chlorpromazine, fluphenazine, paliperidone,
placebo, quetiapine, thioridazine, trifluoperazine, typical
neuroleptic/conventional.
3.2. Overall symptoms
Nine studies reported usable outcome data for mean change
in overall symptoms (see Fig. 3).
There was only one placebo controlled study which
showed no significant difference compared to paliperidone
(N = 1, SMD −0.32, CI −0.71–0.08). Concerning head-to-
head comparisons between different antipsychotics, only
JID: NEUPSY
ARTICLE IN PRESS [m6+;September 19, 2018;17:35]
4 M. Krause et al.

Fig. 1 Plot of direct comparisons – overall symptoms.

The figure shows the direct comparisons of included studies.
The thickness of the line corresponds to the inverse variance of the direct comparisons
The size of the nodes corresponds to the number of trials that study the treatments.

olanzapine was significantly superior compared to haloperi- 3.5. Negative symptoms

dol (N = 2, SMD 0.47, CI 0.10–0.84). There were no sig-
nificant differences between amisulpride and risperidone
(N = 1, SMD, −0.03, CI −0.90–0.84], chlorpromazine and
clozapine (N = 1, SMD, 0.15, CI −0.58–0.88), olanzapine and
risperidone (N = 3, SMD −0.51, CI −1.32–0.30] nor between
quetiapine and risperidone (N = 1, SMD, 0.16, CI −0.29–
0.60).
3.3. Response (authors’ definition)
Five studies reported usable outcome data for the out-
come response (see eFig. 4). Concerning head-to-head com-
parisons between different antipsychotics, only olanzapine
was significantly superior compared to haloperidol (N = 1
OR 2.63, CI 1.04–6.69). There were no significant differ-
ences between amisulpride and risperidone (N = 1 OR 0.94,
CI 0.24–3.71), olanzapine and risperidone (N = 2 OR 0.59, CI
0.14–2.46) nor between quetiapine and risperidone (N = 1,
OR 0.38, CI 0.13–1.06).
3.4. Positive symptoms
Six studies reported usable outcome data for mean change
of positive symptoms (see eFig. 3). There was no placebo-
controlled trial which reported usable outcome data for
positive symptoms. Concerning head-to-head comparisons
between different antipsychotics, there were no significant
differences between amisulpride and risperidone (N = 1,
SMD 0.60, CI −0.29–1.49), chlorpromazine and clozapine
(N = 1, SMD 0.09, CI −0.64–0.82), haloperidol and olanzap-
ine (N = 2, SMD 0.10 CI −0.26–0.47) nor between olanzapine
and risperidone (N = 2 SMD −0.11, CI −0.38–0.16].
Seven studies reported usable outcome data for mean
change of negative symptoms (see eFig. 4). There was
no placebo-controlled trial which reported usable outcome
data for negative symptoms. Concerning head-to-head com-
parisons between different antipsychotics, only olanzapine
was significantly superior compared to haloperidol (N = 2
SMD 0.50, CI 0.02–0.99). There were no significant dif-
ferences between amisulpride and risperidone (N = 1 SMD,
−0.71, CI −1.61–0.18], chlorpromazine and clozapine (N = 1
SMD 0.11, CI −0.62–0.84], nor between olanzapine and
risperidone (N = 3 SMD, −0.88 CI, −2.27–0.52].
3.6. Depressive symptoms
Only four studies reported usable outcome data for mean
change of depressive symptoms (see eFig. 5). There was
no placebo-controlled trial which reported usable outcome
data for depressive symptoms. Concerning head-to-head
comparisons between different antipsychotics, there were
no significant differences between amisulpride and risperi-
done (N = 1 SMD −0.44, CI −1.33–0.44), haloperidol and
olanzapine (N = 1 SMD 0.35, CI −0.05–0.75) nor between
olanzapine versus risperidone (N = 2 SMD 0.10, CI −0.15–
0.36).
3.7. Quality of life
Only two studies reported usable outcome data for the im-
provement in quality of life (see eFig. 7). There was just
one placebo controlled study which showed no significant
difference compared to paliperidone (N = 1, SMD, −0.20

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS [m6+;September 19, 2018;17:35]
Antipsychotic drugs for elderly patients with schizophrenia 5

Fig. 2 Flow-chart of study selection.

The diagram illustrates the process of review and exclusion of studies.

CI −0.59–0.20). Concerning head-to-head comparisons be- 3.9. Prolactin increase

tween different antipsychotics there was only no significant
difference between olanzapine and haloperidol. Only two studies reported usable outcome data for the out-
come prolactin increase (see eFig. 11). Concerning head-to-
head comparisons between different antipsychotics, olan-
zapine showed significantly lower prolactin increase com-
pared to haloperidol (N = 1 SMD −0.66 CI −1.07–−0.25) and
3.8. Social functioning risperidone (N = 1 SMD −1.38 CI −1.80–−0.97).

Only two studies reported usable outcome data for the im-
provement in social functioning (see eFig. 8). There was
just one placebo controlled study which showed virtually 3.10. Weight gain
no difference compared to paliperidone (N = 1 SMD −0.01,
CI −0.41–0.39). Concerning head-to-head comparisons be- Five studies reported usable outcome data for the out-
tween different antipsychotics there was no significant dif- come weight gain (see eFig. 9). There was just one placebo
ference for haloperidol compared to olanzapine. controlled study which showed no difference compared

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS [m6+;September 19, 2018;17:35]
6 M. Krause et al.

Table 1 Characteristics of included studies.

Study Study groups Number of Trial Mean Minimum Mean Duration Year Blinding
participants duration antipsychotic age age ill (years)
(weeks) doses
(mg/days)
Andia et al. (1998) Clozapine 13 3 150 46 68,4 N.a. 1998 OL-RCT
Andia et al. (1998) Haloperidol 13 6 N.a.
Barak et al. (2000) Risperidone 26 72 N.a. 65 N.a. 30 2000 N.a.
Barak et al. (2000) Typical neuroleptic 25 N.a. N.a. 30
Barak et al. (2002) Haloperidol 10 N.a. 7.2 63 69,2 31,4 2002 OL-RCT
Barak et al. (2002) Olanzapine 10 13.1 69,2 39,2
Berman et al. (1995) Haloperidol 10 N.a. N.a. 57 67,4 N.a. 1995 DB-RCT
Berman et al. (1995) Risperidone 10 N.a 67,4 N.a.
Bora (1968) Carphenazine 15 26 N.a. 50 N.a. N.a. 1968 N.a.
Bora (1968) Trifluoperazine 15 N.a N.a. N.a.
Branchey et al. (1978) Fluphenazine 15 8 N.a. 60 67 N.a. 1978 DB-RCT
Branchey et al. (1978) Thioridazine 15 N.a 67 N.a.
Feldman et al. (2003) Olanzapine 20 28 18.8 50 56,9 N.a. 2003 DB-RCT
Feldman et al. (2003) Risperidone 19 8.9 57,2 N.a.
Howanitz et al. (1999) Chlorpromazine 18 12 600 55 68,5 40 1999 DB-RCT
Howanitz et al. (1999) Clozapine 24 300 65 38
Jeste et al. (2003) Olanzapine 89 8 11.1 60 71,4 38 2003 DB-RCT
Jeste et al. (2003) Risperidone 87 1.9 70,9 34,9
Kennedy et al. (2003) Haloperidol 34 6 9.4 60 65,8 N.a. 2003 DB-RCT
Kennedy et al. (2003) Olanzapine 83 11.9 66 N.a.
Kinon et al. (2003) Conventional 143 52 N.a. N.a. N.a. N.a. 2003 N.a.
Kinon et al. (2003) Olanzapine 150 N.a N.a. N.a.
Lacro (2001) Haloperidol 15 12 3.4 N.a. 58 N.a. 2001 DB-RCT
Lacro (2001) Risperidone 12 1.8 58 N.a.
Mintzer et al. (2004) Quetiapine 65 16 237 60 66 N.a. 2004 OL-RCT
Mintzer et al. (2004) Risperidone 27 3.3 67 N.a.
Riedel et al. (2009) Amisulpride 25 6 198 65 72,7 N.a. 2009 DB-RCT
Riedel et al. (2009) Risperidone 13 1.9 72,7 N.a.
Ritchie et al. (2003) Olanzapine 34 4 9.9 58 69,5 N.a. 2003 OL-RCT
Ritchie et al. (2003) Risperidone 32 1.7 69,7 N.a.
Salganik et al. (1998) Clozapine 17 10 N.a. 60 66,6 N.a. 1998 DB-RCT
Salganik et al. (1998) Haloperidol N.a. N.a 66,6 N.a.
Tzimos et al. (2008) Paliperidone 76 6 8.4 65 70 36 2008 DB-RCT
Tzimos et al. (2008) Placebo 38 0 69 31
DB-RCT double blind randomized controlled trial; OL-RCT open label randomized controlled trial; N.a. not available.

to paliperidone (N = SMD 0.00, CI −0.40–0.40). Concern- significant difference compared to olanzapine (N = 1 OR

ing head-to-head comparisons between different antipsy-
chotics, there were no significant differences between
haloperidol and olanzapine (N = 2 SMD −0.98, CI −2.21–
0.25) nor between olanzapine and risperidone (N = 2 SMD
0.42, CI −0.06–0.90).
3.11. Use of at least one antiparkinson
medication
Four studies reported usable outcome data for the out-
come use of at least one antiparkinson medication (see eFig.
12). There was just one placebo controlled study which
showed no significant difference compared to paliperidone
(N = 1 OR 0.58, CI 0.23–1.47). Concerning head-to-head
comparisons between different antipsychotics only olan-
zapine was significantly superior compared to haloperidol
(N = 2 OR 4.81 CI 2.11–10.95). Finally risperidone showed no
1.18 CI 0.50–2.80).
3.12. Dropouts due to any reason
Eight studies reported usable outcome data for the out-
come dropouts due to any reason (see Fig. 4). There was
just one placebo controlled study which showed a nonsignif-
icant superiority for paliperidone (N = 1, OR 0.41, CI 0.16–
1.02). Concerning head-to-head comparisons between dif-
ferent antipsychotics, olanzapine was significantly better
accepted compared to risperidone (N = 3, OR 0.54, CI 0.31–
0.93). There were no significant differences between chlor-
promazine and clozapine (N = 1, OR 1.38, CI 0.17–10.82,
haloperidol and olanzapine (N = 1, OR 1.00, CI 0.17–5.98),
haloperidol and risperidone (N = 1, OR 7.00, CI 0.33–150.06)
nor between quetiapine versus risperidone (N = 1, OR 0.98,
CI 0.37–2.62).

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS [m6+;September 19, 2018;17:35]
Antipsychotic drugs for elderly patients with schizophrenia 7

Fig. 3 Meta-analysis - change of overall symptoms.

CI, confidence interval; SMDs were obtained from a random-effects model assuming a common heterogeneity across all drugs. A
negative value means that the first mentioned drug is favored and vice versa.

3.13. Dropouts due to inefficacy
Five studies reported usable outcome data for the outcome
dropouts due to inefficacy (see eFig. 14). There was just
one placebo controlled study which showed significantly less
dropouts due to inefficacy for paliperidone (N = 1 OR 0.22,
CI 0.05–0.93). Concerning head-to-head comparisons be-
tween different antipsychotics, fairly there were no signifi-
cant differences between haloperidol and olanzapine (N = 1
OR 0.26 0.02–3.06) nor between olanzapine and risperidone
(N = 3 OR 0.84, CI 0.26–2.74).
3.15. Assessment of heterogeneity and small trial
bias
We did not detect significant heterogeneity in any outcome.
As there were very few studies for most of the comparisons,
heterogeneity might not be well estimated. Funnel plots to
detect small trial/publication bias were not meaningful, be-
cause the maximum number of trials available for a compar-
ison was three, not enough to determine asymmetry of the
plots.

3.14. Dropouts due to adverse events
Five studies reported usable outcome data for the outcome
dropouts due to adverse events (see eFig. 13).There was no
significant difference for this outcome.
4. Discussion
To the best of our knowledge this is the first system-
atic review and meta-analysis regarding the effects of an-
tipsychotics in elderly patients with schizophrenia. The
main findings were that olanzapine was significantly more

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS [m6+;September 19, 2018;17:35]
8 M. Krause et al.

Fig. 4 Meta-analysis - dropouts due to any reason (acceptability).

OR, odds ratio; CI, confidence interval; ORs were obtained from a random-effects model assuming a common heterogeneity across
all drugs. A value < 1 means that the first mentioned drug is favored and vice versa.

efficacious compared to haloperidol, based on two trials, of
which one showed no significant effects due to small sample
size (N = 20). Moreover, olanzapine showed significant supe-
riority compared to haloperidol for the reduction of nega-
tive symptoms. It also showed significantly higher response
rates, lower prolactin increase, and less use of antiparkin-
son medication compared to haloperidol. While the only
placebo-controlled study evaluating paliperidone showed no
significant effect for any symptoms of schizophrenia, the
number of dropouts due to inefficacy was significantly lower
in the paliperidone group.
We did not examine differences between second-
generation and first-generation antipsychotics as classes,
because this classification has been replaced with “Neuro-
science Based Nomenclature (NbN)” by societies such as the
European and the American Colleges of Neuropsychophar-
macology (Zohar et al., 2014).
4.1. Comparison to other meta-analyses
How do these results in differential efficacy and side ef-
fects of antipsychotics compare to findings from previous
meta-analytic studies in the general population which often
excluded elderly patients? Compared to a large network
meta-analysis (NMA), do antipsychotics in treatment of
schizophrenia (Leucht et al., 2013) which included 212
RCT’s with data for 43,049 patients in the general pop-
ulation (mean age 38.4), the direction of effects were
generally similar, although the effect sizes differed and al-
though most of our findings were not statistically significant.
As the previous NMA which reported a small superiority for
olanzapine compared to haloperidol (SMD –0.14 CI, –0.21 to
–0.08), we also found a moderate effect favoring olanzapine
in our meta-analysis about elderly patients (SMD −0.47, CI
−0.84 to −0.10). In terms of prolactin increase we found

Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
JID: NEUPSY
ARTICLE IN PRESS
Antipsychotic drugs for elderly patients with schizophrenia
significant differences favoring olanzapine compared
to risperidone (SMD − 1.38, CI − 1.80 to − 0.97] and to
haloperidol (SMD −0.66, CI −1.07 to −0.25). However the
effect sizes for the earlier NMA in the general population
were again smaller: olanzapine vs. risperidone (SMD –1.09,
CI –1.28 to –0.90) and olanzapine vs. haloperidol (SMD –0.56,
CI –0.73 to –0.40). Except of these trend differences in a
few comparisons, the insufficient evidence of the subgroup
was not able to reveal clear differences in the effects of
antipsychotic treatment in elderly patients compared to
the earlier meta-analyses using mostly younger patients. In
particular for those comparisons and outcomes for which
no evidence was available, clinicians therefore need to rely
on the results in the “general” people with schizophrenia,
keeping the specific characteristics of elderly people such
as higher side-effect vulnerability and reduced metabolism
into account.
4.2. General limitations
The present meta-analysis is not without limitations. For
many comparisons only one study was available. This is not,
however, a reason to not systematically present and review
their effect sizes which primarily depend on sample size and
not on number of studies. We systematically reviewed all
the relevant studies available and presented results graph-
ically using standardized statistical procedures utilized in
modern meta-analysis. But it is possible that some com-
parisons which showed non-significant trends might show
significant effects if statistical power was increased if a
larger number of subjects and/or studies were available for
that comparison. The lack of data is particularly important
for some of the side-effect evaluations. In the supplement
eFigs. 10–12 show that data on important side effects like
QTC prolongation, prolactin increase or use of at least one
antiparkinson medication were often not reported in the
original studies. A reason for the low frequency of report-
ing for these outcomes may be due to the fact that most
of the included publications from which we extracted our
data were only subgroup analysis of larger studies, and some
were presented only in abstracts. This is also an important
reason for the high frequency of unclear risk of bias ratings
and small sample sizes.
4.3. Elderly vs geriatric patients
Another important issue is that the included studies used
very heterogeneous definitions of “elderly”. The minimum
age ranged from 46 up to 65 and the mean age from around
57 up to nearly 73. Only three out of eighteen included stud-
ies included patients with a minimum age of sixty-five. How-
ever, in a sample with more uniformly older age patients, all
above age 60 or 65, it is quite possible we might still not find
significant differences in efficacy or side effects of antipsy-
chotic drugs, in elderly patients with schizophrenia. Many
patients these days in their 60’s are in fairly good health
and do not differ in important ways from patients 10 or even
20 years younger. Age itself may not be the most important
differentiating factor. In order to explore whether there are
Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
[m6+;September 19, 2018;17:35]
9
meaningful differences we might have to focus more partic-
ularly on patients who would be classified as geriatric by
recent definitions. The German Society of Geriatrics (DGG),
the German Society of Gerontology and Geriatrics (DGGG),
and the German Group of Geriatric Institutions (BAG) devel-
oped a definition of the geriatric patient (Sieber, 2007) in-
cluding mainly an age of over 70 and a high multimorbidity.
There are not sufficient trials with antipsychotic drugs in pa-
tients with schizophrenia or related psychosis who meet this
more specific geriatric classification, but, as the popula-
tion ages, more patients will be falling into this group. More
studies of antipsychotics in patients with schizophrenia or
related psychosis are needed, in patients who meet inclu-
sion criteria for being truly geriatric as described above
(Sieber, 2007), in order to determine whether there are im-
portant differences in efficacy or side effects in this geri-
atric group who are treated with antipsychotics for these
illness.
Contributors
The authors had full access to all of the data in the study
and take responsibility for the integrity of the data and the
accuracy of the data analysis. SL and MK designed this study.
MK, MH, JS, PR, HR, SB, LB, MS, TA, LR and NP extracted
the data. MK and RS analyzed data. MK and SL wrote the
first draft of the manuscript and RS revised it. All authors
contributed to and have approved the final manuscript.
Declaration of interests
SL has received honoraria for consulting from LB Pharma,
Lundbeck, Otsuka, TEVA, Geodon Richter, Recordati, LTS
Lohmann, and Boehringer Ingelheim; and for lectures from
Janssen, Lilly, Lundbeck, Otsuka, SanofiAventis, and Servier.
MH has received speaker’s honoraria from Janssen and Lund-
beck. Robert Smith is a consultant for Abbot on the devel-
opment of a new medication. The other authors declare no
competing interests.
Acknowledgments
This work was supported by a grant from the German Fed-
eral Ministry of Education and Research (Bundesministerium
für Bildung und Forschung, BMBF, Grant no. FKZ 01KG1508).
We thank Samantha Roberts for help in the literature search
and Leonie Reichelt, Hannah Röder, Susanne Bächer, Lio
Bäckers, Myrto Samara, Thomas Arndt and Natalie Peter for
help in data extraction. We thank all authors of the included
studies, particularly those who sent us additional informa-
tion about their trials.
Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at doi:10.1016/j.euroneuro.
2018.09.007.
JID: NEUPSY
ARTICLE IN PRESS
10
References
Andia, I., Zumarraga, M., Zabalo, M.J., Bulbena, A., Davila, R.,
1998. Differential effect of haloperidol and clozapine on
plasma homovanillic acid in elderly schizophrenic patients
with or without tardive dyskinesia. Biol Psychiatry 43 (1),
20–23.
Andreas, S., Schulz, H., Volkert, J., Dehoust, M., Sehner, S., Sul-
ing, A., Ausín, B., Canuto, A., Crawford, M., Da Ronch, C.,
Grassi, L., Hershkovitz, Y., Muñoz, M., Quirk, A., Rotenstein, O.,
Santos-Olmo, A.B., Shalev, A., Strehle, J., Weber, K., Wegschei-
der, K., Wittchen, H.-U., Härter, M., 2017. Prevalence of men-
tal disorders in elderly people: the European MentDis_ICF65+
study. Brit. J. Psychiatry J. Ment. Sci. 210 (2), 125–131. doi:10.
1192/bjp.bp.115.180463.
Barak, Y., Shamir, E., Weizman, R., 2000. Risperidone compared
with typical neuroleptic treatment in elderly schizophrenic pa-
tients. Int. J. Neuropsychopharmacol. 3, S122 (Abstracts of the
XXIInd CINP Congress, Brussels, Belgium, July 9-13, 2000).
Barak, Y., Shamir, E., Zemishlani, H., Mirecki, I., Toren, P., Weiz-
man, R., 2002. Olanzapine vs. haloperidol in the treatment of
elderly chronic schizophrenia patients. Prog. Neuropsychophar-
macol. Biol. Psychiatry 26 (6), 1199–1202.
Berman, I., Merson, A., Allan, E., Alexis, C., Losonczy, M., 1995.
Effect of risperidone on cognitive performance in elderly
schizophrenic patients: a double-blind comparison study with
haloperidol. Psychopharmacol. Bull. 31 (3), 552.
Bian, Z.X., Li, Y.P., Moher, D., Dagenais, S., Liu, L., Wu, T.X.,
Miao, J.X., Kwan, A.K., Song, L., 2006. Improving the quality of
randomized controlled trials in Chinese herbal medicine, part I:
clinical trial design and methodology. Zhong xi yi jie he xue bao.
J. Chin. Integr. Med. 4 (2), 120–129.
Bora, G., 1968. Comparison of dosage ranges of carphenazine and
trifluoperazine in elderly chronic schizophrenics. Dis. Nerv. Syst.
29 (10), 695–697.
Branchey, M.H., Lee, J.H., Amin, R., Simpson, G.M., 1978. High-
and low-potency neuroleptics in elderly psychiatric patients.
JAMA 239 (18), 1860–1862.
Cohen, J., 2013. Statistical Power Analysis for the Behavioral Sci-
ences. Elsevier Science, Burlington, p. 459.
Colenda, C.C., Mickus, M.A., Marcus, S.C., Tanielian, T.L., Pin-
cus, H.A., 2002. Comparison of adult and geriatric psychi-
atric practice patterns: findings from the American Psychi-
atric Association’s Practice Research Network. Am. J. Geri-
atric Psychiatry Off. J. Am. Assoc. Geriatric Psychiatry 10 (5),
609–617.
DerSimonian, R., Laird, N., 1986. Meta-analysis in clinical trials.
Controll. Clin. Trials 7 (3), 177–188. doi:10.1016/0197-2456(86)
90046-2.
Divine, G.W., Brown, J.T., Frazier, L.M., 1992. The unit of analy-
sis error in studies about physicians’ patient care behavior. J.
General Internal Med. 7 (6), 623–629.
Elbourne, D.R., Altman, D.G., Higgins, J.P.T., Curtin, F., Worthing-
ton, H.V., Vail, A., 2002. Meta-analyses involving cross-over tri-
als: methodological issues. Int. J. Epidemiol. 31 (1), 140–149.
Essali, A., Ali, G., 2012. Antipsychotic drug treatment for elderly
people with late-onset schizophrenia. Cochrane Database Syst.
Rev. 2, CD004162. doi:10.1002/14651858.CD004162.pub2.
Feldman, P.D., Kaiser, C.J., Kennedy, J.S., Sutton, V.K., Tran, P.V.,
Tollefson, G.D., Zhang, F., Breier, A., 2003. Comparison of
risperidone and olanzapine in the control of negative symptoms
of chronic schizophrenia and related psychotic disorders in pa-
tients aged 50 to 65 years. J. Clin. Psychiatry 64 (9), 998–1004.
Furukawa, T.A., Barbui, C., Cipriani, A., Brambilla, P., Watan-
abe, N., 2006. Imputing missing standard deviations in meta-
analyses can provide accurate results. J. Clin. Epidemiol. 59 (1),
7–10. doi:10.1016/j.jclinepi.2005.06.006.
Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
[m6+;September 19, 2018;17:35]
M. Krause et al.
Fusar-Poli, P., Papanastasiou, E., Stahl, D., Rocchetti, M., Carpen-
ter, W., Shergill, S., McGuire, P., 2015. Treatments of nega-
tive symptoms in schizophrenia: meta-analysis of 168 random-
ized placebo-controlled trials. Schizophr. Bull. 41 (4), 892–899.
doi:10.1093/schbul/sbu170.
Higgins, JPT, 2011. Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0. Wiley and Sons, Chichester, UK [up-
dated March 2011].
Howanitz, E., Pardo, M., Smelson, D.A., Engelhart, C., Eisen-
stein, N., Losonczy, M.F., 1999. The efficacy and safety of cloza-
pine versus chlorpromazine in geriatric schizophrenia. Erratum.
J. Clin. Psychiatry 60 (5), 341.
Jeste, D.V., Barak, Y., Madhusoodanan, S., Grossman, F.,
Gharabawi, G., 2003. International multisite double-blind
trial of the atypical antipsychotics risperidone and olan-
zapine in 175 elderly patients with chronic schizophre-
nia. [Erratum appears in Am J Geriatr Psychiatry. 2004
Jan-Feb;12(1): 49]. Am. J. Geriatr. Psychiatry 11 (6),
638–647.
Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative
syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13
(2), 261–276.
Kennedy, J.S., Jeste, D., Kaiser, C.J., Golshan, S., Maguire, G.A.,
Tollefson, G., Sanger, T., Bymaster, F.P., Kinon, B.J., Dossen-
bach, M., Gilmore, J.A., Breier, A., 2003. Olanzapine vs
haloperidol in geriatric schizophrenia: analysis of data from a
double-blind controlled trial. Int. J. Geriatr. Psychiatry 18 (11),
1013–1020.
Kinon, B.J., Stauffer, V.L., Kaiser, C., Kollack-Walker, S., 2003. In-
cidence of presumptive tardive dyskinesia in elderly patients
treated with olanzapine or conventional antipsychotics. In: Pro-
ceedings of the 156th Annual Meeting of the American Psychi-
atric Association. San Francisco CA Nr55.
Lacro, J., 2001. Antipsychotic Treatment in Late Life Schizophrenia.
CRISP Database.
Leucht, S., Cipriani, A., Spineli, L., Mavridis, D., Örey, D.,
Richter, F., Samara, M., Barbui, C., Engel, R.R., Geddes, J.R.,
Kissling, W., Stapf, M.P., Lässig, B., Salanti, G., Davis, J.M.,
2013. Comparative efficacy and tolerability of 15 antipsychotic
drugs in schizophrenia: A multiple-treatments meta-analysis.
Lancet 382 (9896), 951–962.
Leucht, S., Förstl, H., Bäuml, J., 2012. Kurzlehrbuch Psychiatrie
und Psychotherapie [+ campus.thieme.de]. Thieme, Stuttgart,
p. 87 chapter 4, ISBN-10: 3131483814.
Leucht, S., Pitschel-Walz, G., Engel, R.R., Kissling, W., 2002.
Amisulpride, an unusual “atypical” antipsychotic: a meta-
analysis of randomized controlled trials. Am. J. Psychiatry 159
(2), 180–190. doi:10.1176/appi.ajp.159.2.180.
McGrath, J., Saha, S., Chant, D., Welham, J., 2008. Schizophrenia:
A concise overview of incidence, prevalence, and mortality. Epi-
demiol. Rev. 30, 67–76. doi:10.1093/epirev/mxn001.
Mintzer, J.E., Mullen, J.A., Sweitzer, D.E., 2004. A comparison of
extrapyramidal symptoms in older outpatients treated with que-
tiapine or risperidone. Curr. Med. Res. Opin. 20 (9), 1483–1491.
Overall, J.E., 1962. The brief psychiatric rating scale. PR 10 (3),
799. doi:10.2466/PR0.10.3.799-812.
Remington, G., Foussias, G., Fervaha, G., Agid, O., Takeuchi, H.,
Lee, J., Hahn, M., 2016. Treating negative symptoms in
schizophrenia: an update. Curr. Treat. Options Psychiatry 3, 133–
150. doi:10.1007/s40501- 016- 0075- 8.
Riedel, M., Eich, F.X., Moller, H.J., 2009. A pilot study of the safety
and efficacy of amisulpride and risperidone in elderly psychotic
patients. Eur. Psychiatry 24 (3), 149–153. doi:10.1016/j.eurpsy.
2008.10.005.
Ritchie, C.W., Chiu, E., Harrigan, S., Hall, K., Hassett, A., Mac-
farlane, S., Mastwyk, M., O’Connor, D.W., Opie, J., Ames, D.,
2003. The impact upon extra-pyramidal side effects, clinical
JID: NEUPSY
ARTICLE IN PRESS
Antipsychotic drugs for elderly patients with schizophrenia
symptoms and quality of life of a switch from conventional to
atypical antipsychotics (risperidone or olanzapine) in elderly pa-
tients with schizophrenia. Int. J. Geriatr. Psychiatry 18 (5),
432–440.
Salganik, I., Modai, I., Bercovici, B.R., Kutzuk, D., Weizman, A.,
1998. Clozapine vs haloperidol therapy in elderly chronic
schizophrenic inpatients - Preliminary results. A double-blind,
cross-over randomized study. Int. J. Geriatr. Psychopharmacol.
1 (4), 185–187.
Sieber, C.C., 2007. Der ältere Patient–wer ist das? Der. In-
ternist 48 (11), 1192–1194. doi:10.1007/s00108- 007- 1945- 3,
1190.
Leucht, S., Huhn, M., Rothe, P., Schneider, J., Zhu, Y, 2016. Which
are the most important first-generation antipsychotic drugs?
Survey of international schizophrenia experts. In: Abstracts
from the 5th Biennial SIRS Conference - Poster Abstracts. NPJ
Schizophrenia 25.
Please cite this article as: M. Krause et al., Antipsychotic drugs for elderly patients with schizophrenia: A systematic review
and meta-analysis, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.09.007
[m6+;September 19, 2018;17:35]
11
Tzimos, A., Samokhvalov, V., Kramer, M., Ford, L., Gassmann-
Mayer, C., Lim, P., Eerdekens, M., 2008. Safety and tolerabil-
ity of oral paliperidone extended-release tablets in elderly pa-
tients with schizophrenia: a double-blind, placebo-controlled
study with six-month open-label extension. Am. J. Geriatr. Psy-
chiatry 16 (1), 31–43. doi:10.1097/JGP.0b013e31815a3e7a.
United Nations, Department of Economic and Social Affairs, Pop-
ulation Division, 2015. World population ageing, 2015. United
Nations, New York, 1 online resource (164).
Wu, T., Li, Y., Bian, Z., Liu, G., Moher, D., 2009. Randomized trials
published in some Chinese journals: how many are randomized?
Trials 10, 46. doi:10.1186/1745- 6215- 10- 46.
Zohar, J., Nutt, D.J., Kupfer, D.J., Moller, H.-J., Yamawaki, S.,
Spedding, M., Stahl, S.M., 2014. A proposal for an updated neu-
ropsychopharmacological nomenclature. Eur. Neuropsychophar-
macol. J. Eur. Coll. Neuropsychopharmacol. 24 (7), 1005–1014.
doi:10.1016/j.euroneuro.2013.08.004.