IPA Bulletin
Recent Advances - Volume 22, Number 3
By Professor John
O'Brien, Dr. Robert Barber, and Professor Robert Baldwin
Effects of immunotherapy on clinical
symptoms, CSF markers and brain
atrophy in patients with Alzheimer’s
disease?
Readers of the Bulletin will be aware that
the clinical trial of Abeta immunization
in patients with Alzheimer’s disease (AD)
was interrupted because of meningoencephalitis
in a proportion of immunized
patients. But were there any benefits to
using this novel approach to treatment of
neurodegenerative conditions like AD?
Gilman and colleagues have recently published
their findings from the interrupted
trial (Neurology. 2005;64(9):1553-62).
The study was a randomized, multicenter,
placebo-controlled, double-blind
trial (phase IIa) using immunotherapy
AN1792(QS-21) in patients with mild
to moderate Alzheimer disease (AD).
Following reports of meningoencephalitis
6% (18/300), immunization was stopped
though double-blind assessments were
maintained for 12 months. Antibody
response was obtained in 19.7% (59/300)
patients.
No significant differences were found
between antibody responder and placebo
groups for ADAS-Cog, Disability Assessment
for Dementia, Clinical Dementia Rating,
MMSE, or Clinical Global Impression
of Change. There were differences favouring
antibody responders in analyses of the
z-score composite across the neuropsychological
test battery though the magnitude
of this response was small (0.03 +/- 0.37 vs
-0.20 +/- 0.45; p = 0.020). CSF tau was also
decreased in antibody responders (n = 11)
vs placebo subjects (n = 10; p < 0.001). On
the basis of these findings the authors concluded
that the trial provides an indication
that Abeta immunotherapy may be useful
in Alzheimer disease, though the clinical
applicability of this type of immunotherapy
remains in doubt.
As part of the same study, Fox et al
(Neurology. 2005;64(9):1563-72) used serial
MRI measurements of brain, ventricular
and hippocampal volumes to examine the
impact of this immunotherapy. Patients immunized
with AN1792 who were antibody
responders had greater decrease in brain
volume (3.12% vs 1.98% over about 11
months) and ventricular enlargement
(1.10 +/- 0.75 vs 0.48 +/- 0.40%; p < 0.001).
As mentioned neuropsychological test battery
showed differences favoring antibody
responders over placebo. The reason for this
dissociation between brain volume loss and
cognitive function in antibody responders
is unclear but the authors speculate that the
volume changes could result from amyloid
removal and associated cerebral fluid shifts.
Preventing Alzheimer’s disease
with raloxifene?
Yaffe and colleagues undertook a randomized,
placebo-controlled trial in
postmenopausal women with osteoporosis
using raloxifene, a selective estrogen receptor
modulator (Am J Psychiatry 2005;
162(4):683-90). As a secondary outcome,
they also examined whether raloxifene
affected the risk for Alzheimer’s disease.
After 3 years of following up 5,386 women,
5,153 (95.7%) were cognitively normal, 181
(3.4%) had mild cognitive impairment, and
52 (1.0%) had dementia, 36 with Alzheimer’s
disease. They found women receiving
the higher dose of raloxifene (120 mg/day)
had a lower risk of mild cognitive impairment
(relative risk, 0.67) and Alzheimer’s
disease (relative risk=0.52) compared to
those taking placebo.
Statins and Alzheimer’s disease
The role of cholesterol lowering statins
in both the prevention and treatment
of Alzheimer’s disease remains uncertain.
In what the authors considered to be a pilot
study, 67 patients with mild to moderate
Alzheimer’s disease (AD) were randomised
to either placebo or atorvastatin calcium
(80 mg) for 1 year (Sparks et al, Arch
Neurol. 2005;62(5):753-7). Using standardized
primary and secondary outcome
measures for cognition and behaviour, they
found atorvastatin produced significant
improvements in the Alzheimer’s Disease
Assessment Scale-cognitive subscale and the
Geriatric Depression Scale at 6 months. It
also reduced circulating cholesterol levels.
The 12-month data showed a trend for
improvements in the Alzheimer’s Disease
Assessment Scale-cognitive subscale, Clinical
Global Impression of Change Scale,
and Neuropsychiatric Inventory Scale. The
authors concluded that atorvastatin could
offer clinical benefit in the treatment of AD
but their findings need to be substantiated
by a larger multicenter trial.
Cholinesterase inhibitors and their use
in treating problematic behaviours in
Alzheimer’s disease
Herrmann et al (Am J Geriatr Psychiatry.
2005;13(6):527-534) used data pooled
from three studies of galantamine in mild
to moderate Alzheimer’s disease (n=2033)
to complete a meta-analysis of its effects on
behavioural symptoms. In summary, they
found that changes from baseline in the
neuropsychiatric inventory (NPI) scores
were significantly better in the galantaminetreated
subjects than the placebo group,
though the magnitude of the effect size
was small. Symptoms most responsive to
treatment were hallucinations, anxiety,
apathy and aberrant motor behaviours, with
the authors speculating that this group of
symptoms may be specifically responsive to
cholinergic modifying drugs. Importantly,
we know from a separate study that many of
these symptoms have a negative impact on
the quality of life of carers (Shin et al Am J
Geriatr Psychiatry. 2005;13(6):469-474).
Vitamin E and donepezil for the
treatment of mild cognitive impairment
We know subjects with mild cognitive
impairment are at higher risk of
developing Alzheimer’s disease (AD) than
normal elderly people (10-15% per vs 1-2%
per year). So knowing whether there are
treatments that could delay the onset of AD
in this group is clearly an important issue.
Petersen and colleagues set about trying
to answer this question by completing a
large (n=769) three-year double-blind trial
during which subjects with MCI were randomised
to receive either vitamin E (2000
IU daily), donepezil (10mg daily) or placebo
(N Engl J Med. 2005;352(23):2379-88).
During the study 16% of subjects (212/769)
progressed to develop AD.
Overall, they found that although the
donepezil group had a reduced likelihood
of progressing to Alzheimer’s disease during
the first 12 months of the study (P=0.04),
there were no significant differences at the
three-year primary end point. Vitamin E
had no effect at any time point. In a
secondary analysis, the apolipoprotein E
epsilon4 allele had a noticeable modifying
effect such that in subjects with one or
more apolipoprotein E epsilon4 alleles, the
benefit of donepezil compared to placebo
was evident at 12, 24 and 36 months: a
finding not observed between the vitamin E
and placebo groups. However, the authors
conclude that there is insufficient data to
warrant recommending APOE genotyping
people with MCI. They also suggest that
although donepezil treatment may delay the
clinical progression to AD (by 12 months),
as there was no difference after 3 years this
intervention does not appear to modify the
underlying disease mechanism.
Alzheimer’s disease and mortality:
a 15-year epidemiological study.
Understanding the long-term outcome of
Alzheimer’s disease (AD) has implications
for patients, their families and service
development. Ganguli et al (Arch Neurol.
2005;62(5):779-84) followed up a cohort of
1670 adults 65 years and older for 15 years
to determine outcomes in AD, such as the
duration of survival and mortality rates.
The mean (SD) duration of survival was 5.9
(3.7) years and AD was a significant predictor
of mortality. It was associated with an
increased risk of mortality of 40% in the
whole cohort, and population-attributable
risk of mortality of 4.9%. However, there
were differences between men and women,
with an increased mortality risk only observed
among women.
Neuropsychological tests – can they
predict incident Alzheimer’s disease?
Can neuropsychological tests accurately
predict onset of Alzheimer disease?
Tierney and colleagues from Toronto
(Neurology. 2005;64(11):1853-9) followed
up participants of the Canadian Study
of Health and Aging (CSHA) to see
whether they could predict who developed
Alzheimer’s disease after five and 10 years.
Subjects who were free of dementia at baseline
completed a range of neuropsychological
tests. With sensitivities and specificities
of around 70-80% they found three tests
(short delayed verbal recall, animal fluency,
and information) predicted conversion to
Alzheimer’s disease after five years, whilst in
the in the 10-year follow-up study, only one
test (short delayed verbal recall) emerged as
predictive. The authors concluded that these
neuropsychological tests could indeed accurately
predict people who are more likely
to develop Alzheimer disease in the future.
Insights into Alzheimer’s disease
using neuroimaging
Neuroimaging is increasingly being
used to analyse the effects of drug
treatments on patients with dementia. Mega
et al (Arch Neurol. 2005;62(5):721-8) used
fludeoxyglucose F 18 positron emission
tomography (PET) to study the effects of
galantamine in 19 patients with mild to
moderate Alzheimer’s disease (AD). They
compared images before and after treatment
and found for the total group there was an
increase in left caudate metabolism. Comparing
responders and non-responders they
found patients classified as cognitive and
behavioral responders had greater activation
of a striatal-thalamofrontal network with
treatment than patients whose condition
worsened or was unchanged by therapy.
Furthermore, changes in left anterior cingulate
metabolism significantly correlated
with change in the cognitive function,
whereas right cingulate metabolic change
correlated with improvement in depression
and right ventral putamen metabolic
change with improvement in apathy. This
led the authors to conclude that cognitive
and behavioral responders to galantamine
therapy show clinically related improvements
in prefrontal network metabolism
along with thalamic activation.
Are there any neuroimaging changes that
could predict patients that are at highest
risk of developing AD? From the point of
structural imaging, Stoub et al (Neurology.
2005;64(9):1520-4) used repeated volumetric
measures of the entorhinal cortex
and hippocampus to see whether they could
predict the risk of AD. At baseline subjects
were either free of dementia (n=35) or had
a diagnosis of amnestic mild cognitive
impairment (MCI) (n=23). They found
measures of entorhinal atrophy, but not
hippocampal atrophy, were independent
predictors of incident AD, suggesting that
entorhinal cortex atrophy is indeed associated
with risk of Alzheimer disease.
Results from a separate study from France
(Chetelat et al Neuroimage. 2005 Jun 22;
[Epub ahead of print]) overlap with these
findings though found some differences.
Using an imaging technique to measure
grey matter (GM) loss over 18 months in
subjects with mild cognitive impairment
(MCI), they investigated whether there
were differences in patients who convert
to a diagnosis of Alzheimer’s disease and
those who remain stable. Seven of the 18
patients with MCI converted to a diagnosis
of AD and these patients showed greater
GM loss than non-converters at baseline in
the hippocampus, parahippocampal cortex,
and lingual and fusiform gyri and as well
as accelerated atrophy in the hippocampal
area, inferior and middle temporal gyrus,
posterior cingulate, and precuneus. These
findings could be relevant in monitoring
the effects of treatment on brain structure.
Do non-pharmacological interventions
for sleep disturbance in Alzheimer’s
disease work?
In a relative small study (n=36) of patients
with Alzheimer’s disease who were experiencing
sleep problems, researchers from
Washington, USA (McCurry et al, Am
Geriatr Soc. 2005;53(5):793-802) found the
use of behavioural techniques (specifically,
sleep hygiene education, daily walking, and
increased light exposure) were beneficial —
suggesting these techniques can be applied
to help patients and their carers. Clearly,
further clinical trials would be beneficial.
Citalopram for anxiety disorders
Little is known about the treatment of
anxiety disorders in later life. Lenze and
colleagues (American Journal of Psychiatry,
2005; 162: 146-150) report on an eightweek
RCT of citalopram in the treatment
of late life anxiety disorders. Although 791
elderly people were screened, only 47 signed
consent forms and 34 were randomised.
Nevertheless, in spite of the small size, there
was a significant effect size (Hamilton Anxiety
Scale) for citalopram. As to side-effects,
there was little difference in reported
problems in the placebo group but a higher
rate in the citalopram group which, however,
decreased over the course of the study.
On the whole then there was evidence of
efficacy and tolerability for citalopram.
Clearly a larger trial is indicated.
Association between peripheral
inflammatory markers and late-life
depression
Thomas et al (American Journal of Psychiatry,
2005;162:175-7) report on an
increase in interleukin-1 beta in late life depression.
They hypothesise that there would
be higher levels of IL-1 beta in subjects with
major depression and that this would correlate
with depression severity. It did and also,
there were no differences between controls
and sub-syndromal depression (n=20 and
21 respectively). Although a cross-sectional
study it suggests that reported changes in
inflammatory markers such as this may be
state related, although it does not rule out
an independent pathway caused by inflammation.
Ethnicity and depression
Van der Wurff and colleagues (Journal
of Affective Disorders, 2004;83: 33-41)
explored the interesting issue of ethnicity
in migrants. From a sample of older Dutch
elderly people (55+) they interviewed about
300 each of Turkish, Moroccan and Dutch
subjects with the Center for Epidemiologic
Depression Scale (CES-D). The rate of
depression amongst the Dutch was as previously
reported by the same group but there
was a much higher rate for migrants and
especially those of Turkish extraction. The
response rate amongst the Turkish elderly
was relatively low and the study was crosssectional
but nevertheless it is one of the
first to assess an important issue - the effects
of migrant work on mood.
Also linked to ethnicity, Sleath and colleagues
from North Carolina, US (Journal
of American Geriatric Society, 2005;53:
397-404) found that almost a third of
over 2,000 care givers of elderly male US
veterans with dementia were suffering from
depression (CES-D score >9). Although
white care givers were almost twice as likely
as African Americans to have depressive
symptoms the latter were less likely to be
using antidepressants. Of “enabling” factors
only subjective social support related to
antidepressant use. Of “need” factors care
givers who rated the health status as worse
were more likely to be taking anti-anxiety
agents. Overall the antidepressant usage
rate was only 19% of the study. The authors
suggest that many more care givers could
benefit from them.
Strothers and colleagues from Atlanta,
US (Journal of the American Geriatric
Society, 2005;53:456-61) report mixed
news about antidepressant treatment in the
elderly whilst providing another insight
into ethnicity. They used a large Medicaid
database (which should ensure that cost is
not an important component of a decision
to treat). The good news is of increased rates
of antidepressant treatment generally and
with newer antidepressants in particular.
However, patients of an African-American
background were almost twice as likely
as whites not to receive antidepressants.
Clearly a study of this sort cannot provide
explanatory data, but the authors speculate
that factors such as patient preference, support
systems including religious affiliation
and perhaps systemic factors (e.g. Medicaid
providers may be less likely to locate in
minority neighbourhoods) are all possible
factors.
Depression in the oldest old —
don’t give up Stek and colleagues (American Journal
of Psychiatry, 2005;162:178-80), from
Holland, examined the interplay between
depression, loneliness and mortality, using
the Geriatric Depression Scale (n=476). All
subjects were part of the Leiden 85+ Study
meaning that subjects were followed up
from their 85th birthday until mid 2002,
with the end point being survival in days.
In line with other work, the presence of
significant depression had an important
effect on mortality, even after controlling
for other obvious variables. The presence of
perceived loneliness contributed strongly to
the effect of depression on mortality. The
authors conclude that in this group, the oldest
old, the effect of depression on mortality
may occur only when feelings of loneliness
are present. Clearly this would repay further
study and the authors speculate that “motivational
depletion” (“giving up”) may be a
factor.
Hippocampal volume reduction
in depression
In a previous number we mentioned the
role of the hippocampus in depression.
Hickie and colleagues (British Journal of
Psychiatry, 2005;186:197-202) produced
data on 66 patients with depression and 20
controlled participants who underwent a
brain scan and neuropsychological testing.
The age range (28 to 82) included 49 with
an early-onset (aged under 50) and 17
with a late-onset. In this specialist sample,
hippocampal volume reduction was found
in the patient group and correlated with
age, age of onset and reduced cognitive and
memory performance. However hippocampal
volume reduction also occurred in early
onset patients. Depression is a risk factor
for dementia and this finding appears to go
against the idea that depression is simply a
prodromal phase of a late onset dementia.
In this study there were no associations
of hippocampal volume with vascular or
genetic risk factors.
Depression and vascular disease:
exploring the link
Two review articles emphasise the important
but evermore complex issue of the
relationship between cerebrovascular disease
and late-life depression (often referred
to as “vascular depression”). Baldwin (International
Journal of Geriatric Psychiatry,
2005;20:1-11) used a modified conceptual
framework derived from well established
work used to define disease states and their
causes. These include factors such as the
strength, consistency, temporal relationship,
biological gradient along with the plausibility
and coherence of the arguments, in
relation to causality. It is clear that there is
a good body of evidence linking subcortical
brain lesions to depression and that there
is both a gradient and a temporal relationship
— with newer data showing that
lesions tend to progress and are associated
with a worsened prognosis for depression.
Kales et al (American Journal of Geriatric
Psychiatry, 2005;13:88-98) use the more
traditional approach of elucidating biological
pathways which might make an individual
vulnerable, subject to certain other
triggers. Both papers highlight the dif-
ficulty in specifically identifying a “vascular
depression” because of the bi-directional
nature of the link between depression and
vascular disease. Both also highlight the
mixed evidence that commonplace vascular
risk factors, in themselves, are associated
with depression. Elucidating potential
mechanisms could lead to novel treatments
designed to halt vascular damage.
Does decreased depression improve
physical functioning in older adults?
This was the question posed by Callahan
et al (Journal of the American Geriatric
Society, 2005;53:365-73). As part of the
large collaborative care study of depression
in later life (IMPACT) the authors analysed
outcomes for physical items and activities of
daily living. Patients whose mood symptoms
improved also experienced improvement in
physical outcomes. This was independent of
subjective well-being and included practical
tasks such as management of money
and medication. The study lasted for one
year with improved physical outcomes
throughout. Conversely those whose depression
did not improve also experienced no
improvement in their physical functioning.
As the authors point out, even modest improvements
in physical outcome can make
the difference between dependence and independence
and the treatment of depression
seems to be one factor in helping patients
stay the right side of the line.
Medical comorbidity and depression
Leading on from this, a themed edition of
the American Journal of Geriatric Psychiatry
(2005, Volume 13, no. 3) is devoted to
medical comorbidity and includes a number
of papers relevant to depression. A few will
be summarised. Horowitz and colleagues
(American Journal of Geriatric Psychiatry,
2005;13:180-7) established the frequency
of major and minor depression in 584 community
residents seeking help from visual rehabilitation
services. The prevalence of major
depression was 7% and that for subthreshold
depression 27%. These figures are more
akin to similar studies conducted in hospital
settings and rather higher than figures obtained
from other communities and those in
primary care. General demographic variables
such as gender and educational level did not
predict depression but this may have been
masked by the high levels of disability which
was strongly linked. Macular degeneration
was the most frequently reported disorder although
interestingly severity of self-reported
vision was less consistently associated with
depression status. On the whole the risk factors
for major and subthreshold depression
were similar.
Pain and depression
The relationship between pain and depression
is complex and bi-directional.
Carp and colleagues (American Journal of
Geriatric Psychiatry, 2005:13:188-94) explored
the relationship between a measure of
chronic pain, medical burden and depression
outcome, treatment being a combination of
medication and psychotherapy. Although
pain did not adversely predict response to
depression, treated depression did not seem
to relieve pain either, leading the authors
to speculate that this may be connected to
the measure which they used. It probably
supplements a study previously reviewed
(Lin et al JAMA 2003, 290:2428-34) which
emphasised the need for both analgesia
and antidepressant treatment. In the same
themed issue Llorente and colleagues from
Miami, Fla., United States (American Journal
of Geriatric Psychiatry, 2005;13:195-201)
highlighted a relationship between prostatic
cancer (the most common newly-diagnosed
malignancy in men in the United States)
and suicide. Twenty suicides associated with
prostate cancer occurred within 12 months
of diagnosis (of 667 total suicides), amounting
to a four-fold increased risk compared to
the remainder of the same-aged population).
In keeping with the previous study, pain may
be a particular factor and men may at risk
because they try to be stoic.
A good review
For readers who do not want to trawl
through a large number of papers
on late-life mood disorders, an excellent
seminar (basically an extensive review
article) in the Lancet (Alexopoulos, Lancet,
2005;365:1961-9) provides much of the
information one needs in order to be up-to-
date. With 144 references the areas of
diagnosis, aetiology, epidemiology, pathophysiology,
management and outcome are
all covered. The author concludes that all
the available treatments for younger adults
are effective in later life and there are some
special considerations arising from medical
co-morbidity and vascular disease that will
shape treatment strategies. Unfortunately,
in North America at least, public insurance
does not cover the newly emerging innovative
service delivery models such as collaborative
care (eg IMPACT, as above).
John T. O’Brien, Robert Barber, and Robert Baldwin are the
Research Editors of the IPA Bulletin. They welcome readers’ comments via email
(J.T.O'Brien@ncl.ac.uk).
