Progression of brain lesions in migraine.
The Population-Based CAMERA MRI follow-up study

Background:
Migraine is a highly disabling, multifactorial, episodic, neurovascular brain
disorder, affecting 12% of the general population and associated with a 3-7
fold increased risk of cerebral and myocardial infarction. In 2000 we scanned
295 migraineurs and 140 healthy subjects randomly identified from the general
population with MRI and found an increased risk of cerebellar infarcts and
cerebral, cerebellar and pontine white matter hyperintensities in migraineurs
(Dutch CAMERA study: Kruit MC et al; JAMA 2004, Brain 2005, Stroke 2006; NHS
grant 97-108). The risk was greatest in those with aura and higher with
increasing attack frequency, suggesting a causal relationship. Possible
mechanisms for the brain abnormalities include:
i) repeated cerebral hypoperfusion during attacks (accompanying cortical
spreading depression, the underlying mechanism for migraine aura); and
ii) paradoxical cerebral emboli due to the high prevalence (up to 60% of patent
foramen ovale (PFO) in migraineurs. Functional studies in clinic-based
migraineurs have shown sub-clinical cerebellar dysfunction (Sandor P, et al.
Annals Neurol
2001) and increased recruitment of brain areas when executing motor tasks
(Rocca MA, et al. Stroke 2003), suggesting impaired brain function in severe
migraineurs. To establish if repeated migraine attacks may indeed affect brain
structure and function, an unbiased, population-based, longitudinal study is
needed, assessing, within the same cohort, the presence, change and
severity of:
- migraine characteristics;
- MRI brain lesions; and
- brain function.
Positive answers to these questions would call for an earlier and more
agressive prevention of migraine attacks
and associated brain damage.
We propose to:
- rescan the original CAMERA cohort with MRI
- compare the results with the scans from 2000; and
- correlate the findings with results from (follow-up) tests for migraine
characteristics, cardiovascular risk factors,
brain function, and (now newly assessed) presence of PFO.

Hypothesis:
We hypothesize that repeated migraine attacks are associated with progressive
brain damage and impaired brain function as a result of recurrent cerebral
ischemia during recurrent migraine attacks. We predict that migraine
patients, in particular those with aura and high attack frequency, will show an
increase in brain lesion load and impaired brain function when compared to
controls and with examinations done seven years ago in the CAMERA I study.

In the present study we want to demonstrate progression of brain lesions and
associated impairment of brain function in migraineurs vs. non-migraine
controls.

To this extent we have the following objectives:

1) Measure progression of brain lesions of over time in migraine patients and
control subjects over a period of seven years;
2) Establish a correlation between changes in brain lesion load and specific
migraine characteristics (in particular migraine aura, attack frequency, use of
migraine medication) and presence of PFO;
3) Test for impaired cerebellar and cognitive function, and correlate findings
with MRI abnormalities and migraine characteristics.


Divided into categories:

A Evaluation of progression in brain lesions
•Rescanning of CAMERA 1 study population (n=435) using the same MRI systems to
measure progression over time;
•Assessing progression of the lesion load- either by identifying an increase of
the lesion load within already affected patients, or by identifying newly
affected patients;


B Evaluation of contribution of other factors in the cause of lesions in
migraineurs
•Correlating migraine characteristics to difference in lesion load.
•Evaluating the presence of a PFO by means of transcranial Doppler (TCD) and
correlating these data to white matter hyperintensities and infarcts.

C Evaluation of functional consequences
•Test for impaired cerebellar function with a set of non-invasive tests and
correlate findings with MRI abnormalities;
•Test for cognitive function with a cognitive test battery similar to CAMERA 1;
results will be compared with baseline test results.

Final aim is - with answers to these questions - to answer the important
healthcare issue whether migraine should be treated more aggressively and/or
earlier with preventative agents in order to prevent migraine attack-related
brain damage.

CAMERA 2 will be a population-based, longitudinal follow-up study. Participants
of the CAMERA 1 study (n=435) will be re-invited for the proposed follow-up
study. Only previous participants who are still willing and able to give
informed consent will be included (CAMERA 1 inclusion criteria: adults age
30-60 years, signed informed consent, participants of MORGEN-study). In 2000
all participants gave permission to be re-invited for follow-up research.
Subjects with contra-indications for MRI will be excluded. We estimate to
include at least 200 migraine cases and 100 controls.

All participants will undergo the same examinations as was previously performed
including structured telephone interview, MR imaging of the brain, standard
neurological exam, and cognitive testing. This effectively implies a 7 year
follow-up of the same population (longitudinal population-based case-control
study). We will add to this set of measures transcranial doppler (TCD) and a
body sway test (BST). These examinations pose no or minimal risk. All
examinations will be done with the same methods and precautions used in a
standard medical examination, and are supervised by a physician.

In more detail, all participants will undergo:
- a telephone interview, with questions about general health condition, present
and/or past diseases and brain disease risk factors, and (cases) additional
questions about their migraine, focussing on differences compared to 2000. We
will check whether the controls have had migraine attacks in the past 7 years. A
detailed history of migraine characteristics, in particularly aura and attack
frequency, will be recorded.
- re-scanning of the brain using the same 1.5T MRI scanner in Maastricht
(ACS-NT; Philips Medical Systems, Best, The Netherlands) and the same 1.0T MRI
scanner in Doetinchem (Magnetom Harmony; Siemens AG, Erlangen, Germany).
- TCD, to be performed in collaboration with Prof. dr. W.H. Mess
(neurophysiology, AZM, Maastricht), who is an expert in this field, to assess
whether a PFO is present.
- cognitive testing, to be administered by a trained medical students, in
cooperation with Prof.dr. H. Middelkoop (clinical neuropsychology, LUMC,
Leiden). The same test battery as in CAMERA 1 will be used (15 word memory
test, stroop test, word fluency, letter digit substitution test, peg-board
test, depression scale and MMSE). Long and short term memory, concept shifting,
psychomotor speed, organization, visuo-spatial ability, motor speed, and
characteristics of depressive disorder can be characterized by the results of
these tests.
- Cerebellar function tests, in collaboration with prof.dr. C. de Zeeuw
(Neuroscience Institute, EMC Rotterdam) and dr. B. Bloem (Medical director,
Parkinson Center Nijmegen) to measure several cerebellar functions with
non-invasive tests and compare these data between migraineurs and
non-migraineurs. The results of these tests inform us about cerebellar
functioning and the possible influence on it by damage from cerebellar
(ischemic) lesions.

WML-load quantification
An automated image processing method that was recently developed and validated
at the McConnell Brain Imaging Center of the Montreal Neurological Institute
(MNI) under the direction of Dr Alex Zijdenbos for the Age, Gene/Environment
Susceptibility (AGES) study, a collaborative population study conducted by the
National Institute on Aging and the Icelandic Heart Association, will be used
to segment and quantify WML at baseline and follow up to obtain quantitative
data on volume changes of these lesions over time. The elimination of intra-
and inter-rater variability by automatic analysis enables the system to detect
small changes over time, which are not readily detectable through conventional
analysis techniques. The WML quantification is part of a fully automatic
validated "pipeline" image analysis framework, which has been successfully
applied it to a number of large-scale, multicenter studies (>1,000 MRI scans).
This pipeline system is based on robust and sophisticated image processing
algorithms.

Infarct rating
Infarcts will be scored visually, in a qualitative fashion by experienced
readers, because reliable automated methods to automatically segment infarcts
are not available. Baseline scans and follow up scans will be projected next to
each other on a 30 inch screen. Using longitudinal registration and resampling,
corresponding slices will be linked. The software for handling, hanging and
viewing of the images has also been developed for the AGES study. The presence
of infarcts will first be assessed on the follow-up scan. Each detected infarct
will be traced back on the baseline scan, and based on the comparison of the
two scans an assessment will be made for each infarct whether it was already
present on the previous scan, and if so whether it increased in size
(measurement of largest diameter of the infarct in the axial plane) or whether
it remained stable. To avoid any type of bias, two experienced readers will
independently read all MRI images. Before start of the actual rating, readers
will score each 50 selected sets (of baseline and new) MRI examinations for
training purposes; training results will be discussed to steepen learning
curve. During the actual rating, readers are unaware of clinical data. In case
of discrepancies between two readers, a consensus will be reached after
discussion between the two readers. The intra-rater reliability will be
estimated by having 25 pairs of randomly selected scans cycle through the total
sample of scans in a random order so the readers do not know they have already
read the scans.
Infarcts are defined as non-mass parenchymal defects, with a vascular
distribution, isointense to CSF-signal on all sequences, and, when
supratentorial, surrounded by a hyperintense rim on FLAIR and PD images.
Number, location, vascular territory and size of infarcts will be recorded,
including data on increase/decrease in size of existing infarcts and/or
presence of new infarcts. Virchow-Robin spaces will be excluded as infarcts,
based on location, shape, size and absence of a hyperintense border.

All potential participants will be contacted by the study coordinator two weeks
after recieving the invitation. They will be asked to particpate and can ask
questions about the study participantion before deciding to participate. After
signing the informed consent form, an appointment will be made for an interview
by telephone which will take about 20 minutes.

1-3 Weeks after this phonecall the participants will visit the research centre
once. This visit will take a maximum of 3 hours and includes above mentioned
investigations which are all harmless.