Synopsis
Caffeine acute effects on oxygen metabolism are
not well characterized across the brain with MRI. We aim at measuring these in a double-blind,
crossover, placebo-controlled study on sixteen healthy, moderate caffeine
consumers using a dual calibrated fMRI approach and a novel forward estimation model.
Results show spatial variations in OEF0, CBF, CVR,
venous CBV and CMRO2 across grey matter at different levels of
resolution (grey matter, ROI and voxel), in agreement with most of the literature
findings. Therefore we propose this approach as the first viable method to
assess the effects of drugs on brain metabolism with a voxel-wise resolution.Introduction
Caffeine is one of the most common and widely
consumed psycho-active substances. It is a non-selective antagonist of the
adenosine receptors causing stimulation of brain activity, via increased
neuronal firing rates, and reduction of cerebral blood flow (CBF), via
vasoconstriction [1]. Caffeine’s effects can be heterogeneous across the brain,
depending, for example, on regional receptor expression [2]. Previous MRI
studies have shown effects of this drug on brain metabolism, mostly looking at grey
matter changes in CBF [3] and cerebral metabolic rate of oxygen consumption
(CMRO
2), with CMRO
2 having been shown to both increase [4]
and decrease [5]. Recently, methods [6] exploiting fMRI acquisitions during
hypercapnic and hyperoxic respiratory tasks have been developed to estimate
absolute CMRO
2 (an approach known as dual calibrated fMRI or dcFMRI)
potentially allowing brain oxygen metabolism to be mapped in drug studies.
Purpose
We want to show how a dcFMRI approach can be
used to assess the acute effects of caffeine on oxygen extraction fraction (OEF
0),
CBF, cerebrovascular reactivity (CVR), venous cerebral blood volume (CBV) and
CMRO
2 across grey matter.
Methods
Sixteen healthy, moderate caffeine consumers (between
51 and 298 mg/day; 8 females, age = 24.7±5.1) were recruited for a double-blind,
crossover, placebo-controlled study. Each participant was scanned on two separate
days (30.1±18.8 days apart, same time of the day), with each day including two dcFMRI
experiments: one before (“pre” condition) and one about 45 min after (“post”
condition) delivery of an oral caffeine capsule (250 mg). The dcFMRI experiment
consisted of an 18 min dual-gradient echo acquisition (TE
1 = 2.7 ms, TE
2 = 29
ms, 64x64, 3x3x7mm
3, gap = 1 mm, 12 slices) while performing a
respiratory task including three periods of hypercapnia (5% CO
2
balanced air) interleaved with two periods of hyperoxia (50% O
2
balanced air). A spiral k-space acquisition was used along with a PICORE
tagging scheme and a QUIPSS II cut-off at 700ms. These data were analysed using
a novel forward model ([7]), to estimate maps of OEF
0, CBF, CVR, CBV
and CMRO
2. Results from one subject were excluded because of the
poor quality of end tidal traces, leaving N=15 subjects for the following
statistical analysis. Along with grey matter statistics, voxelwise t-tests were
performed on the estimated maps to test the difference of the changes from pre
to post session with placebo and caffeine. Averaged values of each parameter
were also calculated in seven ROIs (caudate nucleus, frontal lobe, insula, occipital
lobe, parietal lobe, putamen and thalamus) to further explore spatial distribution
of the effects.
Results
Plateau
levels of hyperoxia and hypercapnia caused an average increase of approximately
210 mmHg and 11 mmHg in end-tidal O
2 and CO
2 pressure
respectively, compared to baseline (Fig. 1).
Grey
matter values of CBF, CBV and CMRO
2 averaged across the subjects
significantly decreased by 30.4% (±6.1%), 31% (±13.8%) and 18.6% (±11.1%) with
caffeine, compared to a non-significant variation of 1.3% (±7.3%), -2% (±23.5%)
and 0.7% (±21.3%) respectively with placebo (Fig. 2, B-D-E). A significant
difference is also found for OEF
0 with an increase of 15.6% (±18.9%),
compared to 1.9% (±19.7%), while for CVR the increase of 17.4% (±27.7%) with
caffeine is not significantly different from the increase of 7.9% (±23.1%) with
placebo (Fig. 2, A-C). T-test maps show a widespread significant reduction compared
to the pre-dose scan with caffeine compared to placebo for CBF and CBV, while
more localized decrease for CMRO
2 and increase for OEF
0
and CVR (Fig. 3). The ROI analysis supplies a more complete picture of the
distribution of the changes in metabolism (Fig. 4), highlighting the different contribution
of CBF and OEF
0 to oxygen consumption across the brain.
Conclusions
Results show significant differences in response
to caffeine compared to placebo, with directions of the effects consistent with
what is expected from previous findings as regards CBF and OEF
0 ([1],[4],[5])
and with part of the MRI literature as regards CMRO
2 [4]. Statistics
on the estimated maps show the spatial distributions of the response to
caffeine both at a ROI and voxelwise resolution, allowing us to localize areas
of significant effects. We therefore propose the dcFMRI approach using our
novel forward model ([7]), as the first viable MRI method to assess the effects
of drugs on brain metabolism with a voxel-wise resolution.
Acknowledgements
We thank the UK Engineering and Physical Sciences Research Council and Cardiff University President's Scholarships for supporting this work.References
[1] Fredholm et al., Pharma. rev.,
83-133 (1999);
[2] Laurienti et al.,
AJNRm 1607–1611 (2003);
[3] Chen et al., Neuroimage, 647-52
(2009);
[4] Griffeth et al., Neuroimage,
809-816 (2011);
[5] Xu. et al, Neuroimage, 39,-47,
(2015);
[6] Wise et al., Neuroimage, 135-147
(2013);
[7] Germuska et al., ISMRM meeting
2015, abstract 2131;