Keywords
Alzheimer’s disease - psychostimulant - systematic review and meta-analysis - apathy
- cognition impairment
Introduction
Alzheimer’s disease (AD) is a common neurodegenerative disorder. Recently, Chicago
health and aging population study showed that the 2010 census standardized prevalence
of dementia caused by AD was 14.5% (95% confidence interval [CI]=13.7–15.3), and the
annual incidence rate was 2.3% (1.7–2.9) [1]. The symptoms of AD are classified into cognitive impairment and behavioral disturbances
[2]. Behavioral disturbances include various psychiatric symptoms, such as depression,
apathy, psychosis, anxiety, agitation, and sleep disturbances [2]. A recent systematic review and meta-analysis showed that memantine treatment is
beneficial for psychosis, agitation, and sleep disturbances in patients with AD [3]. However, apathy is the most common neurobehavioral symptom associated with AD,
but currently approved anti-dementia drugs do not improve this symptom [4].
The pathophysiology of apathy includes impairment of dopaminergic neurotransmission
in the brain areas, such as the ventral tegmental area [5]
[6]. Prolonged exposure to amyloid oligomers decreases the release of glutamate and
gamma-aminobutylic acid, reducing the possibility of dopamine release in the prefrontal
cortex and hippocampus [5]. The progressive decrease of glutamate release from the prefrontal cortex reduces
the stimulus for dopamine release in the nucleus accumbens, resulting in apathy [5]. Patients with AD and apathy show a blunted subjective response to dextroamphetamine
challenge [7], suggesting that the use of psychostimulants increasing dopamine levels in the brain
may improve apathy. A recent double-blind, randomized, placebo-controlled trial (DBRPCT)
of methylphenidate (n=60), a psychostimulant, showed that this drug improved the apathy
evaluation scale-clinician version (AES-C) [8] score in patients with AD as compared with a placebo [9]. However, another DBRPCT of methylphenidate (n=60), which used the apathy evaluation
scale-informant (AES-I) [10], did not show this effect in patients with AD [11]. Thus, the efficacy of methylphenidate remains inconclusive ([Table 1]). The low statistical power (insufficient sample size) of these studies might make
them difficult to accurately estimate the efficacy of methylphenidate. A meta-analysis
can increase the statistical power of group comparisons and overcome the limitations
of sample size in underpowered studies [12]. A recent meta-analysis reported that methylphenidate was superior to placebo in
the improvement of apathy scales score [4]. When different studies use different scales, the Cochrane handbook recommends using
random-effects models and standardized mean difference (SMD) analysis [12]. Our meta-analysis aimed to fill the gap in the literature in terms of the efficacy
and safety of psychostimulants for the treatment of patients with AD. Therefore, we
conducted a comprehensive systematic review and random-effect model meta-analysis
(using SMD for continuous outcomes and risk ratios (RRs) for dichotomous outcomes
as the response measures). This study aimed to produce conclusive evidence for the
efficacy (improvement of apathy, cognitive impairment, activity of daily living, and
burden of caregiving) and safety (discontinuation rate and incidence of individual
adverse events) of a pooled psychostimulant group in patients with AD. We conducted
this systematic review and meta-analysis by combining 2 psychostimulants (methylphenidate
and modafinil) to overcome the limitations of sample size in underpowered studies.
Table 1 Characteristics of included double-blind, randomized, placebo-controlled trials.
|
(1) Study,
(2) Country,
(3) Sponsorship
|
Total n
|
Methods
(1) Study design
(2) Duration
|
Patients
(1) Diagnosis
(2) Inclusion criteria
(3) Study defined disease severity
|
Intervention
|
N
|
Age (mean±SD), years
|
Male (%)
|
Race (%)
|
Apathy scales at baseline (mean±SD)
|
Efficacy outcomesa
|
|
Frakey 2012, USA, industry
|
23
|
(1) DBRPCT
(2) 8w
|
(1) AD, NINCDS-ADRDA
(2) clinically elevated symptoms of apathy based on FrSBe scale, stable dose of ChEI
for 30 days
(3) mild to moderate
|
MOD 200mg/d
|
11
|
75.3±8.34
|
NR
|
NR
|
FrSBe apathy: 95.6±10.8
|
MOD = PLA: FrSBe apathy, IADL, DAFS, Zarit
|
|
PLA
|
11
|
79.4±7.62
|
NR
|
NR
|
FrSBe apathy: 88.9±12.0
|
|
Herrmann 2008, Canada, non industry
|
13
|
(1) DBRPCCOT
(2) 5w (2w treatment phases with a 1w placebo washout between phases)
|
(1) AD, NINCDS-ADRDA
(2) age >55 y, MMSE>10, NPI apathy>1, stable dose of ChEI for at least 3 m
(3) mild to moderate
|
MET 20mg/d
|
13
|
77.9±7.8
|
46.2
|
NR
|
AES: 48.3±11.0
|
MET > PLA: AES-I, NPI apathy
MET = PLA: NPI total, MMSE
|
|
PLA
|
12
|
|
Padala 2017, USA, non industry
|
60
|
(1) DBRPCT
(2) 12w
|
(1) AD, DSM-IV
(2) MMSE>18, AES-C>40, stable dose of ChEI for at least 4 m or antidepressants for at least
2 m
(3) mild
|
MET 20mg/d
|
30
|
77.0±7.5
|
100
|
Non-Hispanic Caucasian 93.3
|
AES-C: 51.8±7.1
|
MET > PLA: AES-C, IADL
MET = PLA: MMSE, Zarit
|
|
30
|
76.2±8.5
|
100
|
Non-Hispanic Caucasian 100
|
AES-C: 47.9±5.7
|
|
Rosenberg 2013, USA and Canada, non industry
|
60
|
(1) DBRPCT
(2) 6w
|
(1) AD, NINCDS-ADRDA
(2) MMSE>10, clinical stability as judged by the local investigator, clinically significant
apathy for at least 4 w (NPI apathy frequency of “often” or greater and an apathy
severity of “moderate” or “marked”), stable dose for the prior 3 m, if treated with
a SSRI
(3) mild to moderate
|
MET 20mg/d
|
29
|
78.0±8
|
41
|
Non-Hispanic Caucasian 97
|
AES: 50±13
|
MET > PLA: NPI apathy
MET = PLA: AES-I
|
|
PLA
|
31
|
75.0±9
|
35
|
Non-Hispanic Caucasian 87
|
AES: 51±11
|
Methods
This meta-analysis was performed according to the preferred reporting items for systematic
reviews and meta-analyses guidelines [13]. The DBRPCTs using psychostimulants for AD were selected. Double-blind, randomized,
placebo-controlled, crossover trials (DBRPCCOT) were included to increase the sample
size for the meta-analysis. A systematic literature search was conducted according
to the following aspects: patient (AD), intervention (psychostimulants), comparator
(placebo), and outcomes (efficacy and safety outcomes). The review has been registered
with PROSPERO (http://www.crd.york.ac.uk/PROSPERO/. CRD42018085983).
Search strategy
To identify relevant studies, 2 authors (T.K. and K.S.) independently searched databases,
such as MEDLINE, Cochrane library, and Scopus, without language restrictions from
the date of inception of these databases to January 6, 2019 using the following keywords:
(random*) AND (stimulant OR psychostimulant OR methylphenidate OR cathinone OR methcathinone
OR cocaine OR dexmethylphenidate OR amphetamine OR methamphetamine OR 3,4-methylenedioxymethamphetamine
OR dextroamphetamine OR lisdexamfetamine OR atomoxetine OR modafinil OR armodafinil
OR dexamphetamine OR bupropion OR mazindol OR selegiline) AND (Alzheimer*). The authors
also searched the following to ensure the comprehensive inclusion of randomized controlled
trials and to minimize the possibility of publication bias: clinicaltrials.gov (http://clinicaltrials.gov/),
ISRCTN registry (https://www.isrctn.com/), and international clinical trials registry
platform (http://www.who.int/ictrp/en/). The authors independently evaluated the inclusion/exclusion
criteria and selected the relevant studies. The references of the included articles
and reviews were also searched for the citations of additional relevant published
and unpublished studies, including conference abstracts.
Data sources, studies sections, and data extraction
The outcomes of our study were apathy (primary) scale score (i. e., frontal systems
behavior scale apathy [14] from 1 study [15], AES-I from 2 studies [11]
[16], and AES-C from 1 study [9]), mini-mental state examination (MMSE) score [17], instrumental activities of daily living scale (IADL) score [18], Zarit burden interview [19] score, all-cause discontinuation rate, discontinuation due to adverse events, and
incidence of individual adverse event. For evaluating apathy, 1 study [15] used frontal systems behavior scale apathy, 2 studies [11]
[16] used both AES-I and neuropsychiatric inventory (NPI) apathy score [20], and another study [9] used AES-C. There were 2 studies using the NPI apathy score. However, because the
results of these 2 studies were consistent ([Table 1]), we did not perform a meta-analysis using only the data of NPI apathy score. Where
possible, an intention-to-treat or a full analysis set population was used. Although
period 1 data was not available (before crossover) for crossover studies, we used
those data for meta-analysis. If the data required for meta-analysis were missing,
the investigators or the industries of the relevant study were contacted and asked
to provide the unpublished data. Moreover, we extracted data from previous systematic
reviews and meta-analyses [4].
Data analysis
The meta-analysis was conducted using the review manager software (version 5.3 for
Windows; http://tech.cochrane.org/revman), and a random-effects model was selected
because of the potential heterogeneity across several studies. Continuous outcomes
were analyzed using SMD with 95% CIs. Lower MMSE and IADL scale scores indicated a
higher level of impairment or more severe symptoms; hence, the algebraic sign of the
numerical scores was reversed for these scales. Dichotomous outcomes were presented
as RRs. The methodological quality of the selected trials was assessed according to
the risk of bias criteria in the Cochrane handbook for systematic reviews of interventions
[12]. Study heterogeneity was evaluated using the heterogeneity statistic (I
2
), considering I
2
≥ 50% to reflect considerable heterogeneity [21]. A sensitivity analysis was conducted for primary outcomes with considerable heterogeneity,
methylphenidate studies vs. modafinil study, [1] and DBRPCTs vs. DBRPCCOT [2]. Because a funnel plot is generally used only if 10 or more studies are included
in the meta-analysis, we did not utilize this plot for exploring potential publication
bias [12].
Results
Of the 617 studies initially identified by searching the literature, 407 were excluded
after reviewing the titles and abstracts. Reviewing the full text resulted in the
exclusion of 1 systematic review and meta-analysis study [4] (Supplementary Fig. 1S). No additional studies were retrieved further from the clinical trial registries.
Risk of bias
Three of 4 studies were DBRPCTs, and one was DBRPCCOT [16] (Supplementary Fig. 2S). Although this DBRPCCOT did not report the data obtained before crossover, we used
crossover data for the meta-analysis. The primary outcome of all the studies was improvement
in the apathy scale score. One of 4 studies was sponsored by the industry [15]. Two of 4 studies did not have any high risk of bias [9]
[11].
Study selection and characteristics
We identified 4 randomized trials that compared psychostimulants and placebo in 156
patients with AD ([Table 1]). All the studies were published in English and were conducted in the USA and/or
Canada. The study duration ranged 2–12 weeks. Sample sizes ranged from 13–60. All
the studies used standardized diagnostic criteria, including the National Institute
of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease
and Related Disorders Association (3 studies) or the Diagnostic and Statistical Manual
of Mental Disorders Fourth Edition (1 study). The mean age of the patients was 76.8
years.
Results of the efficacy outcomes in individual studies included in the current systematic
review and meta-analysis
A study by Frakey (2012) has reported that modafinil was not superior to placebo in
the improvement of the frontal systems behavior scale apathy score, IADL score, and
Zarit burden interview score [15].
A study by Herrmann (2008) has shown that although methylphenidate was superior to
placebo in the improvement of the AES-I and NPI apathy scores, no significant differences
were observed in the MMSE score between the groups [16].
Padala (2017) has reported that although methylphenidate was superior to placebo in
the improvement of the AES-C and IADL scores, no significant differences were observed
in the MMSE and Zarit burden interview score between the groups [9].
In 2013, Rosenberg has shown that methylphenidate was superior to placebo in the improvement
of the NPI apathy score, but no significant differences were observed in the AES-I
score between the groups [11].
Synthesized findings
Combined psychostimulants were superior to placebos in the improvement of the apathy
scale score (SMD=−0.63 (−1.22, −0.04), p=0.04, I2=68%, based on all included studies; ([Fig. 1-1]) and MMSE score (SMD=−0.58 (−1.14, −0.02), I2=61%, p=0.04, based on the 3 methylphenidate studies; ([Fig. 1-2]). However, no significant differences were observed with respect to other outcomes
between the treatment groups (SupplementaryFigs. 3S–8S).
Fig. 1 1 Forest plot of the apathy scale score. 2 Forest plot of the Mini-Mental State Examination
score. (period) 95% CIs: 95% confidence intervals, IV: inverse variance, M-H: Mantel-Haenszel,
MET: methylphenidate, MOD: modafinil, SD: standard deviation, Std mean difference:
standardized mean difference
Because we detected a considerable heterogeneity for the primary outcome, we conducted
2 sensitivity analyses. When the data of 1 modafinil study were excluded from the
meta-analysis for the improvement of the apathy scales score, methylphenidate was
found to be superior to placebo (SMD=−0.82 (−1.43, −0.20), p=0.009, I2=66%). When the data of 1 DBRPCCOT were excluded from the meta-analysis for the improvement
of the apathy scale score, combined psychostimulants were not superior to placebo
(SMD=−0.63 (−1.41, 0.16), p=0.12, I2=79%).
Discussion
We performed a comprehensive systematic review and meta-analysis to obtain robust
evidence of the efficacy and safety of the combined psychostimulants in patients with
AD. Methylphenidate was found to improve apathy and cognitive impairment in patients
with AD. It is a drug approved by the U.S. Food and Drug Administration [22] for the treatment of attention deficit hyperactivity disorder and narcolepsy. Methylphenidate
blocks the reuptake of norepinephrine and dopamine into the presynaptic neuron and
increases the release of these monoamines into the extraneuronal space [22]. The pathophysiology of apathy and cognitive impairment includes hypofunction of
dopamine neurons in the brain [5]
[6]. Methylphenidate can improve these symptoms by restoring the normal function of
dopamine neurons. Although the study duration included in the current meta-analysis
was short, no significant differences were observed in discontinuation rate as well
as incidence of at least 1 adverse event between psychostimulants and placebo. However,
the meta-analysis did not include other safety outcomes, as insufficient data was
available on these outcomes.
Because we detected considerable heterogeneity in the primary outcome, we conducted
2 sensitivity analyses; however, these did not reveal any confounding factors. The
considerable heterogeneity might have been observed because of small sample sizes.
Because we did not utilize funnel plot for exploring potential publication bias, our
study results might include a publication bias. Moreover, because the number of RCTs
and patients included in our meta-analysis was small, we cannot rule out a “small
study effect,” in which smaller studies tend to show larger treatment effects than
larger studies [23].
Limitations
First, the number of studies and patients included in this meta-analysis are limited.
Second, because all the included studies had short trial durations, we could not determine
whether psychostimulants would have long-term effects on apathy and cognitive impairment.
Although the use of psychostimulants poses a risk of drug dependence, cardiovascular
disease, psychiatric symptoms, such as psychosis and mania, as well as seizures [22], the current systematic review and meta-analysis did not evaluate the association
between psychostimulants and these risks in patients with AD. Moreover, we did not
perform a meta-analysis in terms of safety outcomes other than discontinuation rate
and incidence of at least 1 adverse event. Third, because a funnel plot is generally
used only if 10 or more studies are included in the meta-analysis, we did not utilize
such method for exploring potential publication bias [12].
Conclusions
Our results suggest that methylphenidate is effective in treating apathy and cognitive
impairment in patients with AD. However, the number of patients and studies included
in the current systematic review and meta-analysis was limited. Moreover, the duration
of the included studies was short. Therefore, we considered that a long-term study
with larger sample size must be conducted to obtain robust results.
Contributors
All authors had full access to all study data and are responsible for the integrity
of the data and the accuracy of any data analysis. All authors drafted the final manuscript.
