INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int J Geriatr Psychiatry 2001; 16: 513517.

DOI: 10.1002/gps.383

Olfactory identication is impaired in clinic-based patients

with vascular dementia and senile dementia
of Alzheimer type
Alison J. Gray1*, Vince Staples2, Karen Murren2, Avtar Dhariwal2 and Peter Bentham3
1

Senior Registrar, Worcestershire Community Healthcare NHS Trust, UK

Senior Registrar, South Birmingham Mental Health NHS Trust, UK
3
Consultant in Old Age Psychiatry, South Birmingham Mental Health NHS Trust, UK
2

SUMMARY
Aims It is now well established that there are abnormalities in the sense of smell in patients suffering from Alzheimer's
disease (AD). They have both raised olfactory thresholds and impaired odour identication. The situation in vascular dementia is unclear. We used the University of Pennsylvania Smell Identication Test (UPSIT), a 40-item, forced choice, cued,
`scratch-and- sniff' test, to examine olfactory identication in vascular dementia and to determine whether it would differentiate the disorder from AD and normal elderly.
Methods We investigated three matched subject groups: 13 people having a Cambridge Examination for Mental Disorders
in the Elderly (CAMDEX) diagnosis of denite senile dementia of Alzheimer type, 13 having a CAMDEX diagnosis of denite vascular dementia and 13 non-cognitively impaired controls. The subjects were then tested with the UPSIT in their own
home by an independent blind researcher to see if the test could distinguish the different diagnostic groups in this setting.
Results The median UPSIT score was 30 (out of a maximum of 40) for controls, 12 for the vascular group and 15 for the AD
group. The difference was signicant ( p 0.05) between both demented groups and the normal controls. Similarly there was a
signicant difference in the UPSIT score between the AD group and controls ( p 0.001) and between the vascular dementia
group and controls ( p 0.001), but there was no signicant difference between the AD group and the vascular dementia
group. The UPSIT score correlated strongly with the degree of cognitive impairment as measured by the CAMCOG
(rs 0.683, p 0.01)
Conclusions Patients with vascular dementia had a similar degree of olfactory impairment to those with AD. The UPSIT
successfully differentiated between dementia patients and normal elderly British subjects tested in their own homes. The
UPSIT did not differentiate between those with AD and vascular dementia. Copyright # 2001 John Wiley & Sons, Ltd.
key words Alzheimer's disease; vascular dementia; olfaction; olfactory discrimination; diagnosis; differential diagnosis

INTRODUCTION
Olfactory impairments in Alzheimer's disease (AD)
were rst reported by Waldton in 1974 and it is now
well established that those with AD have abnormal
olfactory function. Impairments have been shown on
tests of olfactory identication, olfactory recognition
memory and olfactory threshold (Mesholam et al.,

*Correspondence to: Dr A. J. Gray, St Anne's Orchard, Link top,

Malvern, Worcs, WR14 IEZ, UK. E-mail: a.j.gray@tesco.net
Copyright # 2001 John Wiley & Sons, Ltd.

1998). The odour detection threshold is higher in

those with more advanced Alzheimer-type dementia
(Murphy et al., 1990). Performance on olfactory identication tests is positively correlated with cognitive
abilities (Serby et al., 1991). In patients with questionable dementia, those with an apolipoprotein E
(ApoE) 4 allele have been shown to have more
impaired odour detection abilities than those without
this allele (Bacon et al., 1998). The combination of
being anosmic at testing and carrying an ApoE 4
allele increases the risk of future cognitive decline.
Normal elderly people who were anosmic at baseline
Received 5 April 2000
Accepted 10 October 2000

514

a. j. gray et al.

and carried at least one ApoE 4 allele have been

shown to have almost ve times greater risk of developing cognitive decline over 2 years compared with
controls without these risk factors (Graves et al.,
1999). A simple three-item test of olfactory identication will differentiate between those with Alzheimer's
disease and those with major depression (Solomon
et al., 1998). Olfactory decits are also prominent
in other neurodegenerative disorders, particularly Parkinson's disease (Doty et al., 1988). The situation is
less clear in vascular dementia, with olfactory
abnormalities being reported in patients with multiinfarct dementia, but not so pronounced as those
found in AD (Knupfer and Spiegel, 1986).
The aims of our study were to determine whether
patients with vascular dementia have signicant
impairment of olfactory identication and whether a
simple olfactory identication test would correctly
distinguish between clinic-based patients with AD,
vascular dementia and normal elderly.

METHOD
Potential subjects were identied from case records of
patients under the care of the Mental Health Services
for Older Adults in South Birmingham, UK. Informed
consent was obtained from the patient and, where
appropriate, a carer. Patients were then interviewed
with the Cambridge Examination for Mental Disorders in the Elderly (CAMDEX) (Roth et al., 1988).
Two patient groups were selected, those with a CAMDEX diagnosis of `denite senile dementia of Alzheimer-type' and those with a diagnosis of `denite
vascular dementia'. Patients were excluded if they
suffered from disorders, or were taking medication
known to impair olfaction or if their dementia was
severe. Subjects with vascular dementia were then
individually matched with AD subjects by: age
(within 10 years), gender, smoking status and Clinical
Dementia Rating (CDR) stage (Hughes et al., 1982).
Normal controls were recruited from spouses of
patients from a mixed clinic population. The controls
had no signicant evidence of cognitive impairment
on the CAMDEX. They were again individually
matched with the vascular subjects on the above criteria except CDR stage.
Subjects were then tested at home, on another occasion, by a second investigator blind to their diagnosis,
with the University of Pennsylvania Smell Identication Test (UPSIT) (Doty et al., 1984). The UPSIT has
been widely used to study olfactory identication deficits in many different groups, correlates well with
Copyright # 2001 John Wiley & Sons, Ltd.

other more time-consuming olfactory tests (Doty

et al., 1984) and is reliable on retesting (Doty et al.,
1985).
The UPSIT is a 40-item test presented as four booklets, each page of which contains a `scratch-and-sniff'
microencapsulated smell panel. Above each strip of
odorant is a four-item multiple-choice question to
aid identication. The items follow a standard format.
An example is: `This odour smells most like: a) chocolate; b) banana; c) onion; or d) fruit punch.' This is a
forced choice procedure; for the test to be valid, subjects must make a choice, even if they smell nothing.
The tests were administered in a standardized format.
Firstly a short explanation was given and then the
odour was released from the test strip and immediately the booklet was given to the subject to smell.
As the subjects smelled the booklet, the four choices
were read out twice. Subjects were also presented
with the choices in large print, with the word
`gasoline' translated into `petrol'. The UPSIT was
developed in the USA and hence some of the test
itemsfor example `root beer' and `bubble gum'
are likely to be unfamiliar to elderly British residents.
It was explained to the subjects that the test is American and that they may be unfamiliar with some items.
They were also given a standardized explanation of
the unfamiliar items as they occurred, e.g. `Dill
pickle, that's vinegar and herbs.'
RESULTS
As expected following the individual matching process, the three groups overall were very closely
matched on demographic, cognitive and behavioural
variables (Table 1). Thirteen subjects with vascular
dementia were matched with 13 AD subjects and 13
normal controls. In each of the groups there were four
males and nine females and each group contained one
smoker. The mean age was 75.4 years (range 71.6
79.2 years) for those with AD, 79.2 years (range
76.182.4 years) for those with vascular dementia
and 75.6 years (range 72.879.2 years) for the control
group. This reects the older average age of those
with vascular dementia but was not signicantly different between the groups, possibly due to small sample size (ANOVA: f 2.046, p 0.144). In each of
the dementia groups there were 11 subjects with
mild dementia (CDR stage 1) and two with
moderate dementia (CDR stage 2). The median
CAMCOG score was 75 for the AD group and 62
for the vascular dementia group and 94 for normal
controls (KruskalWallis 2 18.8, p 000). The
Int J Geriatr Psychiatry 2001; 16: 513517.

olfactory identication in dementia

515

Table 1. Controls, vascular dementia (VaD) and Alzheimer's disease (AD) groups: background demographic data, performance on
dementia measures and UPSIT scores with statistical analysis
Three-group
comparison
(two-tailed)

Two-group
comparison
(two-tailed)

AD group

VaD group

Control group

Gender
male
female

4
9

4
9

4
9

Smoking status
smoker
non-smoker

1
12

1
12

1
12

75.4
(71.679.2)

79.2
(76.182.4)

75.6
(72.879.2)

ANOVA:
f 2.046, p 0.144

75
(5379)

62
(3279)

94
(82102)

KruskalWallis
2 18.8, p 000

Wilcoxon paired test:

AD vs VaD:
Z 1.650, p ns
AD vs controls:
Z 3.185, p 0.001
VaD vs controls:
Z 3.182,
p 0.001

0
11
2

0
11
2

13
0
0

15 (1027)

12 (827)

30 (1634)

KruskalWallis
2 21.987, p 0.001

Wilcoxon paired text:

AD vs VaD:
Z 0.980, p ns
AD vs controls:
Z 3.189, p 0.001
VaD vs controls:
Z 3.182, p 0.001

0.403
0.172

0.374
0.208

0.683z
< 0.01z

Mean (range) age, years

Median (range) CAMCOG*

CDR#
stage 0
stage 1
stage 2
Median (range) UPSITy

Correlation between
UPSIT and CAMCOG
rs
p

0.200
0.512

*Cambridge Cognitive Examination (maximum score 107).

#
Clinical Dementia Rating.
y
University of Pennsylvania Smell Identication Test (maximum score 40).
z
All groups combined.

difference was not signicant between the dementia

groups (Table 1).
The median UPSIT score was 15 for the AD group,
12 for the vascular group and 30 for normal controls
(Figure 1). The difference was signicant (Kruskal
Wallis, 2 21.987, df 2, p 0.001). Post hoc comparisons showed a signicant difference between the
AD group and controls (Wilcoxon paired:
Z 3.189, p 0.001). Similarly, the difference
between vascular and control groups was signicant
(Wilcoxon paired: Z 3.182, p 0.001). There
was, however, no signicant difference between the
two patient groups (Wilcoxon paired: Z 0.980,
p 0.327).
Copyright # 2001 John Wiley & Sons, Ltd.

The UPSIT score strongly correlated with cognitive

performance measured on the CAMCOG (Spearman's correlation coefcient (rs) 0.683, p 0.01),
but was not signicant for the groups taken separately.
A score of 18 on the UPSIT differentiated between
those with dementia and normal controls with a sensitivity of 80.8% (95% condence interval (CI)
6193%) and a specicity of 92.3% (95% CI
64100%).
CONCLUSIONS
As in previous studies, our patients suffering from
clinically diagnosed AD had marked impairment in
Int J Geriatr Psychiatry 2001; 16: 513517.

a. j. gray et al.

516

Figure 1. UPSIT score (median and 95% condence interval) by diagnostic group.

their ability to identify odours. In contrast to previous

ndings, however, patients with vascular dementia
performed at a similar level to those with AD. The
impairment of olfactory discrimination in AD is probably explained by the known propensity of the pathological changes to affect structures involved in
olfaction (Esiri and Wilcock, 1984). It is less clear
why the vascular dementia patients were impaired
to a similar extent, as the distribution of the cerebral
pathological changes in this condition are more varied
and may not involve the olfactory pathways.
It is possible that some of the observed impairment
of odour identication may be due to the cognitive
demands of the task, particularly as there is a signicant correlation between UPSIT and CAMCOG
scores for the combined groups. Cognitive factors,
however, are unlikely to account fully for the
impaired ability to identify odours. Vollmecke and
Doty (1985) used the Picture Identication Test
(PIT) to demonstrate that odour identication decits
in Korsakoff psychosis were not a result of memory
impairment. The PIT is identical in content and format to the UPSIT except that the stimuli are linedrawn pictures rather than odorants. Morgan et al.
(1995) investigated the impact of lexical functioning
Copyright # 2001 John Wiley & Sons, Ltd.

on odour identication, showing that odour identication on the UPSIT is poorer than picture identication
on the PIT in patients with probable and questionable
AD. Similarly they found that odour identication
remained poor even when the lexical demands of
the task were eliminated. Positive correlations
between UPSIT and CAMCOG scores would also
be expected if cognitive decline and progressive
impairment of olfactory abilities were parallel but largely independent processes. Our ndings may support
a view that patients with clinically diagnosed vascular
dementia frequently have a signicant admixture of
Alzheimer-type pathological changes (Kalaria and
Ballard, 1999). The UPSIT was designed and validated in the USA and contains some culturally bound
items which may explain the slightly lower scores in
the normal controls, but not the marked impairment
seen in the vascular dementia group. With minimal
modication the test was readily applicable to an
elderly British population.
The UPSIT is easy to administer, requires little
training and is well tolerated by patients. Other
advantages of using this test include the portability
of the test itself (if required it can be posted) and
the stability of the reagents. It costs around $30 per
Int J Geriatr Psychiatry 2001; 16: 513517.

olfactory identication in dementia

test. The demonstrated sensitivity and specicity were
comparable to those of other diagnostic tests for
dementia and the UPSIT may be clinically useful in
diagnosing mild AD and vascular dementia; however,
it appears not to have any utility in differentiating
between the conditions. Further research should
investigate the clinical utility of the UPSIT in differentiating early AD and vascular dementia from
normal ageing.
ACKNOWLEDGEMENTS
Thanks to Dr Stuart Cumella and Dr David Battin for
support in the early stages. The UPSIT can be purchased from Sensonics Inc., PO Box 112, Haddon
Heights, NJ 08035, USA (http://www.smelltest.com).

KEY POINTS
 Odour identication in vascular dementia is
abnormal to a similar extent to that found in AD.
 The UPSIT can distinguish those with vascular
dementia or AD from normal controls in a
British population.

REFERENCES
Bacon AW, Bondi MW, Salmon DP, Murphy C. 1998. Very early
changes in olfactory functioning due to Alzheimer's disease
and the role of apolipoprotein E in olfaction. Ann NY Acad Sci
855: 723731.
Doty RL, Shaman P, Dann M. 1984. Development of the University
of Pennsylvania Smell Identication Test: a standardised microencapsulated test of olfactory function. Physiol Behav 32:
489502.

Copyright # 2001 John Wiley & Sons, Ltd.

517

Doty RL, Newhouse MG, Azzalina JD. 1985. Internal consistency

and short-term test-retest reliability of the University of
Pennsylvania Smell Identication Test. Chem Senses 10:
297300.
Doty RL, Deems D, Stellar S. 1988. Olfactory dysfunction in
Parkinson's disease: a general decit unrelated to neurologic signs,
disease stage or disease duration. Neurology 38: 12371244.
Esiri M, Wilcock G. 1984. The olfactory bulbs in Alzheimer's disease. J Neuro Neurosurg Psychiatry 47: 5660.
Graves AB, Bowen JD, Rajaram L et al. 1999. Impaired olfaction as
a marker for cognitive decline: interaction with apolipoprotein E
4 status. Neurology 53: 14801487.
Hughes CP, Berg L, Danziger WL, Coben LA, Martin RL. 1982. A
new clinical scale for the staging of dementia. Br J Psychiatry
140: 566572.
Kalaria RN, Ballard C. 1999. Overlap between the pathology of
Alzheimer's disease and vascular dementia. Alzheimer Dis Assoc
Disord 13(Suppl. 3): S115123.
Knupfer L, Spiegel R. 1986. Differences in olfactory test
performance between normal aged, Alzheimer and vascular
dementia individuals. Int J Geriat Psychiatry 1: 314.
Mesholam RL, Moberg PJ, Mahr R, Doty R. 1988. Olfaction in neurodegenerative disease. Arch Neurol 55: 8490.
Morgan CD, Nordin S, Murphy C. 1995. Odor identication as an
early marker for Alzheimer's disease: impact of lexical functioning and detection sensitivity. J Clin Exp Neuropsychol 17:
793803.
Murphy C, Gilmore MM, Seery CS et al. 1990. Olfactory thresholds are associated with degree of dementia in Alzheimer's disease. Neurobiol Aging 11: 465469.
Roth M, Tym E, Mountjoy CQ, Huppert FA, Hendrie H, Verma S,
Goddard R. 1988. CAMDEX: The Cambridge Examination for
Mental Disorders in the Elderly. Cambridge University Press:
Cambridge.
Serby M, Larson P, Kalkstein D. 1991. The nature and course of
olfactory decits in Alzheimer's disease. Am J Psychiatry 148:
357359.
Solomon GS, Petrie WM, Hart JR, Brackin HB. 1998. Olfactory
dysfunction discriminates Alzheimer's disease from major
depression. J Neuropsychiatry 10: 6467.
Vollmecke TA, Doty RL. 1985. Development of the Picture Identication Test (PIT): a research companion to the University of
Pennsylvania Smell Identication Test. Chem Senses 10:
413414.
Waldton S. 1974. Clinical observations of impaired cranial nerve
function in senile dementia. Acta Psychiatr Scand 50: 539547.

Int J Geriatr Psychiatry 2001; 16: 513517.

Copyright of International Journal of Geriatric Psychiatry is the property of John Wiley & Sons, Inc. and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.