Sniffing out Parkinson’s disease

Summary

Olfactory deficits in Parkinson’s disease (PD) were first empirically documented in 1975 by
Ansari and Johnson. Over the ensuing years it has become clear that most PD patients
have olfactory disturbances that are not restricted to a single functional modality. Even in
early stage, untreated PD patients, deficits in olfactory function have been demonstrated,
which is supported by recent neuropathological studies demonstrating that the olfactory
bulb and anterior olfactory nucleus structures may be among the induction sites of PD
pathology. In later pathological stages, the olfactory bulb and tract are among the brain
regions where Lewy bodies and Lewy neurites, the characteristic neuropathological
features of PD, are particularly abundant. Since impairments in the sense of smell may
even precede the development of overt motor symptoms, olfactory testing could prove
valuable in establishing an early diagnosis of PD when other clinical (motor) symptoms are
not apparent yet. Also in the early clinical motor stages of PD, olfactory testing may be
useful as a diagnostic tool, both for distinguishing between PD patients and controls, and
in differentiating between PD and other neurodegenerative disorders. Furthermore, the
pathophysiology underlying the olfactory deficits in PD is far from being elucidated.
The following research questions were addressed in this thesis:
• What is the prevalence and nature of impairments in the different specific olfactory
modalities in PD and how do they relate to other (motor and non-motor) disease
characteristics?
• Which (combination of) olfactory test(s) is best in discriminating PD patients from
control subjects?
• Is it possible to explore the neurophysiological basis of olfactory (dys)function by
means of magnetoencephalography (MEG) in healthy controls and PD patients?
Prevalence and nature of olfactory deficits in PD
The “Sniffin’ Sticks” is a multimodal olfactory test battery that can be used to assess three
different aspects of olfactory function: odour identification, discrimination and detection,
each consisting of 16 items.
In chapter 1, we provided age-specific normative values for the Dutch population (over 45
years of age) for the two culture-dependent components of the “Sniffin’ Sticks” test
battery: odour identification and odour discrimination. In chapter 3, we used these agedependent
normative values to study the prevalence of deficits on the odour
identification and discrimination task in a large population of Dutch PD patients from two
university medical centres. The prevalence of an olfactory deficit in PD patients on any of
the two tasks in this study was 73%. In chapter 2, we assessed the prevalence of olfactory
deficits (odour identification, discrimination and detection threshold) in PD in a large
sample of PD patients from three populations in Australia, Germany, and the Netherlands.
When we applied age-independent criteria for hyposmia, only 3.3% out of a total of 400
PD patients were normosmic. However, when applying age-specific criteria, as we did for
the Dutch cohort in chapter 3, 25.5% of patients were normosmic. From these data we
concluded that, apparently, a significant minority of PD patients does not suffer from
olfactory dysfunction.
The results in chapter 3 further demonstrate an impairment in odour identification in 65%
of PD patients relative to the performance of controls, and an impairment in odour
discrimination in 42% of patients. The results described in chapter 5 indicate that PD
patients have a slight impairment of odour recognition memory that appears to be fully
accounted for by an increase in odour detection threshold. Taken together, these findings
argue against the notion that the olfactory impairments in PD would be based on a single
common underlying deficit, such as an increased odour detection threshold, but suggest
that olfactory dysfunction in PD entails a disturbance of multiple, but not all, olfactory
modalities.
Relationship between olfactory (dys)function and other disease characteristics
Since approximately 25% of PD patients do not appear to have olfactory deficits (see
above), olfactory function might contribute to the phenotypic characterization of PD
patients. Therefore, we wanted to determine the relationship between the different
olfactory modalities and other PD characteristics.
The results described in chapter 3 show that odour identification performance in PD is
related to age and sex, but independent of disease duration or stage. By contrast, odour
discrimination performance was found to decrease with disease duration in PD.
Chapter 4 addresses the relationship between olfactory impairment and other aspects of
phenotypical heterogeneity among PD patients. Apart from the above-mentioned
association between odour discrimination deficits and disease duration, there were no
other significant correlations between olfactory function and motor or (other) non-motor
symptoms in PD, such as cognitive status, psychiatric complications, sleep or autonomic
function. Moreover, there were no significant differences in olfactory test scores (either
measured as a combined test score or each of three olfactory modalities separate)
between patients with different motor phenotypes (tremor-dominant, akinetic-rigid,
postural instability gait difficulty or mixed (chapters 2 and 4).
Diagnostic value of olfactory testing in PD
The results of the studies described in chapters 2 and 3 show that odour identification is
more frequently impaired in PD than odour discrimination and odour detection, and that
an odour identification test allows a better differentiation between patients and controls.
Odour recognition memory, was not independently impaired in PD (chapter 5), and is
therefore not useful as a diagnostic tool to differentiate between PD patients and control
subjects.
In chapter 6, we used extended versions of the odour identification and discrimination
parts of the “Sniffin’ Sticks” and found that adding more items within a single olfactory
modality does not improve the diagnostic accuracy of these tests. By contrast, combining
different olfactory modalities did increase diagnostic accuracy. A combination of an odour
identification and a detection threshold task turned out to be the best in differentiating
between PD patients and control subjects.
Neurophysiological studies of olfactory function
In chapter 7, we determined the number of chemosensory stimuli needed to obtain an
optimal signal-to-noise (S/N) ratio for studying olfactory event-related responses by
means of an olfactometer and electroencephalography (EEG) in healthy controls. The S/N
ratio of olfactory and trigeminal event-related potentials significantly improved up to 60-
80 stimuli, mainly due to a reduction of the noise level. However, in a pilot study involving
both healthy controls and PD patients, applying our EEG results to MEG, we were unable
to obtain consistent olfactory event-related magnetic fields (unpublished observations).
Therefore, we changed focus towards time-series analyses of MEG data instead, as a
means to gain more insight in the neurophysiological aspects of olfactory information
processing in healthy controls and the pathophysiology of olfactory dysfunction in PD.
Chapter 8 describes the results of a study in which we were able to show for the first time
that time-series analysis of MEG data, including spectral power and synchronization
likelihood (a general measure of functional connectivity between brain areas), can be used
to detect odour-induced changes in brain activity in healthy subjects. In addition, we
found differences in odour-induced changes in brain activity between PD patients and
controls using analysis of functional connectivity, but not of spectral power. These
differences in functional connectivity may reflect abnormal olfactory information
processing in PD patients that leads to the clinically observed olfactory impairments.
General discussion
In the general discussion, the data presented in the various chapters of this thesis were
combined and a consideration of the potential implications as well as future research
perspectives was provided. The most striking observations from the first two sections of
this thesis are A) that apparently approximately 25% of PD patients do not suffer from
olfactory dysfunction, B) that the impairment of olfactory function in PD entails a
disturbance of multiple, but not all, olfactory modalities, and C) that a combination of an
odour detection threshold test and an identification test is the best in distinguishing PD
patients from controls. Furthermore, differential characteristics of the odour identification
and discrimination deficits in PD suggest that these olfactory modalities involve at least
partly differential components of the olfactory information processing system.
From the last section, we can conclude that time-series analysis of MEG data is a suitable
method to study odour-induced changes in brain activity. In addition, differences in odourinduced
functional connectivity were found between PD patients and controls. The results
obtained may be used in future olfactory neuroimaging studies to further investigate the
pathophysiology of olfactory dysfunction in PD, in particular moving beyond the mere
administration of odorants to the use of more complex tasks, such as odour identification
or discrimination.