A ketogenic drink improves brain energy and some measures of cognition in mild cognitive impairment

Abstract

Introduction: Unlike for glucose, uptake of the brain’s main alternative fuel, ketones, remains normal in mild cognitive impairment (MCI). Ketogenic medium chain triglycerides (kMCTs) could improve cognition in MCI by providing the brain with more fuel.
Methods: Fifty-two subjects with MCI were blindly randomized to 30 g/day of kMCTor matching placebo. Brain ketone and glucose metabolism (quantified by positron emission tomography; primary
outcome) and cognitive performance (secondary outcome) were assessed at baseline and 6 months later.
Results: Brain ketone metabolism increased by 230% for subjects on the kMCT (P <.001) whereas brain glucose uptake remained unchanged. Measures of episodic memory, language, executive function, and processing speed improved on the kMCT versus baseline. Increased brain ketone uptake was positively related to several cognitive measures. Seventy-five percent of participants completed the intervention.
Discussion: A dose of 30 g/day of kMCT taken for 6 months bypasses a significant part of the brain glucose deficit and improves several cognitive outcomes in MCI

Full text

Featured Article
A ketogenic drink improves brain energy and some measures
of cognition in mild cognitive impairment
M
elanie Fortier
a,
*, Christian-Alexandre Castellano
a
, Etienne Croteau
a,b
, Francis Langlois
c
,
Christian Bocti
a,d
,Val

erie St-Pierre
a
, Camille Vandenberghe
a
, Micha€
el Bernier
a,b
,
Maggie Roy
a,b,e
, Maxime Descoteaux
e
, Kevin Whittingstall
f,g
, Martin Lepage
g,h,i,j
,

Eric E. Turcotte
g,h,i,j
, Tamas Fulop
a,d
, Stephen C. Cunnane
a,b,d
a
Research Center on Aging, CIUSSS de l’Estrie – CHUS, Sherbrooke, Quebec, Canada
b
Department of Pharmacology and Physiology, Universit
e de Sherbrooke, Sherbrooke, Quebec, Canada
c
CIUSSS de l’Estrie – CHUS, Sherbrooke, Quebec, Canada
d
Department of Medicine, Universite de Sherbrooke, Sherbrooke, Quebec, Canada
e
Department of Computer Science, Universit
e de Sherbrooke, Sherbrooke, Quebec, Canada
f
Department of Radiology, Universit
e de Sherbrooke, Sherbrooke, Quebec, Canada
g
CR-CHUS, CIUSSS de l’Estrie – CHUS, Sherbrooke, Quebec, Canada
h
Sherbrooke Molecular Imaging Center, Universite de Sherbrooke, Sherbrooke, Quebec, Canada
i
Department of Nuclear Medicine, Universit
e de Sherbrooke, Sherbrooke, QC, Canada
j
Department of Radiobiology, Universit
e de Sherbrooke, Sherbrooke, QC, Canada
Abstract Introduction: Unlike for glucose, uptake of the brain’s main alternative fuel, ketones, remains
normal in mild cognitive impairment (MCI). Ketogenic medium chain triglycerides (kMCTs) could
improve cognition in MCI by providing the brain with more fuel.
Methods: Fifty-two subjects with MCI were blindly randomized to 30 g/day of kMCT or matching pla-
cebo. Brain ketone and glucose metabolism (quantified by positron emission tomography; primary
outcome) and cognitive performance (secondary outcome) were assessed at baseline and 6 months later.
Results: Brain ketone metabolism increased by 230% for subjects on the kMCT (P,.001) whereas
brain glucose uptake remained unchanged. Measures of episodic memory, language, executive func-
tion, and processing speed improved on the kMCT versus baseline. Increased brain ketone uptake was
positively related to several cognitive measures. Seventy-five percent of participants completed the
intervention.
Discussion: A dose of 30 g/day of kMCT taken for 6 months bypasses a significant part of the brain
glucose deficit and improves several cognitive outcomes in MCI.
Ó2019 the Alzheimer’s Association. Published by Elsevier Inc. All rights reserved.
Keywords: Acetoacetate; Alzheimer’s disease; Beta-hydroxybutyrate; Decanoic acid; Fluorodeoxyglucose; Glucose; Ketone;
Medium chain triglyceride; Mild cognitive impairment; Octanoic acid; PET imaging
1. Introduction
Several conditions that increase the risk of Alzheimer’s
disease (AD) are associated with a regional deficit in brain
glucose uptake on the order of 10%. These conditions
include carrying the Presenilin-1 mutation, presence of one
or two alleles of apolipoprotein E4 (APOE4), family history
of AD, insulin resistance, or being 65 years old [1].
Conflict of interest/Disclosure: S.C.C. has done consulting for or
received honoraria from Bulletproof, Keto-Products, Accera, Nestl
e, Nis-
shin Oillio, and Pruvit. Abitec Corporation provided the MCT for this proj-
ect. Nestl
e has funded some MCT research of S.C.C.’s group. S.C.C. has
recently formed a company, SENOTEC Inc, to develop ketogenic products.
The other authors have no conflicts of interest.
*Corresponding author. Tel.: 11-819-780-2220 x 45252; Fax.: 11-819-
829-7141.
E-mail address: melanie.fortier2@usherbrooke.ca
https://doi.org/10.1016/j.jalz.2018.12.017
1552-5260/Ó2019 the Alzheimer’s Association. Published by Elsevier Inc. All rights reserved.
Alzheimer’s & Dementia -(2019) 1-10

Because the brain glucose deficit in these conditions is pre-
sent before the onset of any cognitive deficit, by definition, it
is presymptomatic, so may be contributing to deteriorating
brain structure and function associated with the onset of
AD [2–7].
The brain’s main fuel is glucose but when plasma glucose
declines for at least 12 hours, which is long enough to
deplete glycogen stores, the brain also readily uses an alter-
native fuel—ketone bodies (or simply ketones: acetoacetate
[AcAc] and beta-hydroxybutyrate [BHB]). In long-term
fasting, ketones can supply 60% of the brain’s energy re-
quirements [8,9], so they are the brain’s most important
replacement fuel for glucose. When brain glucose uptake
is decreased, basal brain ketone uptake is still normal not
only in cognitively healthy older people, but also in mild
cognitive impairment (MCI) and in mild to moderate AD
compared with healthy younger adults [10–13]. Plasma
ketones normally contribute to ,5% of the brain’s energy
requirements, but when they are available in moderately
increased amounts they actually displace glucose as a
brain fuel because ketones are preferentially taken up over
glucose by the brain [14,15]. In response to consuming a
drink providing 30 g/day of ketogenic medium chain
triglycerides (kMCTs) for one month, brain ketone uptake
increased in patients with AD as it would in cognitively
normal adults [16]. Because the brain energy deficit in
MCI and AD is specific to glucose, some degree of brain en-
ergy rescue by ketones appears to be the mechanism by
which cognitive outcomes improve with ketogenic interven-
tions in both MCI and AD [17–21].
Dietary supplementation with kMCTs is a simpler and
more convenient method to moderately increase plasma ke-
tones than dietary energy or severe carbohydrate restriction
[1]. Unlike long-chain fatty acids, the shorter chain length of
kMCT allows them to reach the liver directly via the portal
vein, and to cross the mitochondrial inner membrane without
carnitine-dependent transport. Hence, kMCTs are more
rapidly
b
-oxidized than long-chain fatty acids [22,23]
thereby permitting them to be ketogenic. The 8-carbon
MCT, tricaprylin, is more ketogenic than the 10-carbon tri-
caprin or coconut oil [24], so we refer here to kMCT as
one that contains at least 50% tricaprylin.
Our goal was to address two questions about kMCTs in
MCI: (1) If cognitive improvement occurs with a kMCT
drink, is it a function of the increase in ketone or energy
availability to the brain? (2) Is a kMCT drink well enough
tolerated to make it a feasible strategy to improve cognition
in older people? The primary objective of the Brain ENErgy
Fitness, Imaging and Cognition (BENEFIC) trial was there-
fore to assess whether global or regional metabolic rate of
AcAc and glucose in the brain, measured by positron emis-
sion tomography (PET) with the carbon-11 (
11
C) AcAc and
[
18
F]-fluorodeoxyglucose (
18
F-FDG) tracers, would increase
in MCI when consuming a kMCT drink for 6 months. Our
secondary objectives were to assess whether a kMCT drink
changes (1) global brain energy supply; (2) performance on a
neurocognitive battery; (3) cognitive outcomes relative to
increased delivery of ketones to the brain; and (4) whether
the kMCT drink was well tolerated. Exploratory objectives
included assessing whether regional brain volumes, cortical
thickness, functional connectivity, or cerebral blood flow
change after kMCT in MCI.
2. Methods
2.1. Participants
The BENEFIC trial was conducted with the informed
written consent of all the participants and was approved
by our institutional ethics committee (CIUSSS de l’Estrie–
CHUS, Sherbrooke, Quebec, Canada). It is registered
at ClinicalTrials.gov with identification number
NCT02551419 under the title “Proof of Mechanism of a
New Ketogenic Supplement Using Dual Tracer PET.” Inclu-
sion criteria were male or female aged 55 years and the
presence of MCI [25]. Criteria for MCI were (1) subjective
memory complaint plus objective cognitive impairment in
one or more domains compared with appropriate normative
data (1.5 standard deviation less than the mean); (2) a
Montreal Cognitive Assessment (MoCA) score of 18 to 26
of 30 or a Mini-Mental State Examination (MMSE) score
of 24 to 27 of 30; (3) absence of depression (General Depres-
sion Scale score ,10/30 [26]); and (4) full autonomy for
daily living based on a score of 15 of 24 on the instru-
mental activities of daily living score (functional autonomy
measurement system [27]). Exclusion criteria included diag-
nosis of a major cognitive disorder according to criteria in
the Fifth Edition of the Diagnostic and Statistical Manual
of Mental Disorders [28], use of an acetylcholinesterase in-
hibitor, major depression or history of alcohol or substance
abuse within the past 2 years, smoking, uncontrolled dia-
betes (fasting plasma glucose .7 mM or glycated hemoglo-
bin .6.5%), overt evidence of heart, liver or renal disease,
vitamin B12 deficiency, uncontrolled hypertension, dyslipi-
demia, or thyroid disease, inability to lie down without mov-
ing for 60 minutes (for the brain imaging), or the presence of
implanted metal objects or devices contraindicated for mag-
netic resonance imaging (MRI). Screening tests for all par-
ticipants were reviewed by a collaborating physician
before enrollment.
2.2. Experimental design
Eligible participants were assigned to the active (kMCT)
or placebo treatment using a randomization sequence
(Excel 2010, Microsoft, Redmond, WA) with 1:1 allocation
and six consecutive blocks of 10 participants and then
scheduled for a dual tracer brain PET scan, an MRI, and
a neurocognitive battery. At enrollment, participants
received their first months’ supply of the bottled drink
and a daily logbook. They returned monthly to meet the
study coordinator to have a blood sample drawn and to
receive their next months’ supply of the drink. A second
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-102

and final dual tracer brain PET scan, an MRI, and a cogni-
tive assessment were scheduled during the final week of the
sixth (final) month of intervention.
2.3. Neurocognitive battery
General cognitive status was estimated with the MMSE
[29] and the MoCA [30]. Episodic memory was assessed
by the French version of the 16-item free and cued word
learning and recall test (Rappel Libre/Rappel Indic
e [RL/
RI-16]) [31] and the Brief Visual Memory Test-Revised
[32]. The Trail Making, Stroop Color and Word Interference
(Stroop), and Verbal Fluency (VF) tests from the Delis-
Kaplan Executive Function System [33] and Digit Symbol
Substitution tests from the Wechsler Adult Intelligence
Scale [34] provided information on executive function,
attention, and processing speed, respectively. The Boston
Naming Test (BNT) [35] was used for language ability. A
table of normative scores for each test was used to determine
a Z-score for each subtest and to calculate composite
Z-scores [33,34,36,37]. Tests used for each composite
Z-score are specified in Table 1.
2.4. Ketogenic MCT and placebo drinks
The active kMCT drink was an emulsion containing 12%
Captex 355 (60% caprylic acid, 40% capric acid; Abitec
Corp, Columbus, OH) in lactose-free skim milk. The drink
provided 30 g kMCT in 250 mL bottles (Nalgene, New
York) and was prepared under aseptic conditions at the dairy
pilot plant at Universit
e Laval (Quebec City, Quebec, Can-
ada) using our proprietary process. The placebo drink
contained refined, bleached, winterized, and deodorized
high-oleic acid sunflower oil as the nonketogenic lipid and
was also prepared under aseptic conditions. It was provided
to the participants in the same 250 mL bottles and was
Table 1
Raw scores on the cognitive tests
Cognitive tests
Placebo Active
DPlacebo
versus active
Pvalue
y
PRE POST
Pvalue*
PRE POST
Pvalue*Mean SD Mean SD Mean SD Mean SD
Episodic memory
RL/RI16—Trial 1 Free Recall (/16) 6.4 2.5 6.6 2.4 .638 6.3 2.0 7.6 2.7 .013 .370
RL/RI16—Total Free Recall (/48) 23.3 6.5 22.7 7.8 .844 23.5 6.9 25.4 7.2 .169 .563
RL/RI16—Total Recall (/48)
z
42.7 3.6 40.1 7.1 .232 43.4 4.9 42.3 5.2 .130 .751
RL/RI16—Delayed Free Recall (/16) 9.0 3.4 8.6 3.1 .529 9.8 3.0 9.6 3.1 .703 .908
RL/RI16—Delayed Total Recall (/16)
z
14.8 1.1 13.5 2.5 .026 14.8 1.7 13.9 2.3 .081 .954
BVMT-R—Trial 1 (/12) 3.1 1.9 4.2 2.2 .051 2.6 1.8 3.9 2.4 .027 .908
BVMT-R—Total (/36)
z
13.4 5.5 15.8 6.9 .098 13.8 6.2 16.6 8.2 .066 .795
BVMT-R—Delayed Recall (/12)
z
4.8 2.4 5.6 2.8 .221 5.9 2.4 6.1 3.1 .688 .435
Composite Z-score
z
20.98 0.74 21.08 1.08 .673 20.87 0.79 20.69 0.98 .446 .499
Executive function
Trail Making—Switching
z
(s) 150 55 152 71 .877 136 60 133 57 .705 .885
Stroop—inhibition
z
(s) 93 36 92 39 .629 87 27 83 25 .443 .686
Stroop—inhibition/switching
z
(s) 112 48 118 64 .673 93 29 88 25 .351 .191
VF—letter (total correct)
z
28.3 5.9 27.0 7.0 .477 29.3 10.5 27.9 11.9 .208 .863
VF—categories (total correct)
z
31.5 7.0 29.4 8.3 .047 30.8 6.8 31.1 6.3 .777 .075
VF—switching (total correct)
z
9.8 1.9 9.2 2.9 .423 10.7 3.6 10.6 2.8 .981 .603
VF—switching accuracy
z
7.2 2.5 7.2 3.1 .979 9.6 4.4 8.5 3.0 .294 .370
Composite Z-Score
z
20.72 0.82 20.72 0.97 .763 20.33 0.98 20.32 0.81 .862 .954
Attention and processing speed
Trail Making—Visual Scanning (s) 29 8 33 9 .022 31 11 31 8 .983 .154
Trail Making—number sequencing (s) 60 27 63 33 .809 55 31 44 21 .043 .385
Trail Making—letter sequencing (s) 55 21 59 21 .144 59 30 62 42 .983 .325
Trail Making—motor speed
z
(s) 47 24 41 16 .226 42 29 34 13 .420 .773
Stroop—color naming
z
(s) 35 7 36 7 .087 37 10 36 12 .617 .085
Stroop—reading
z
(s) 26 5 27 6 .659 28 9 28 7 1.000 .402
Digit symbol substitution test
z
(/133) 44.9 11.9 47.2 14.9 .312 47.9 13.6 46.9 15.2 .452 .708
Composite Z-score
z
20.02 0.57 20.08 0.71 .840 20.06 0.87 0.08 0.79 .171 .191
Language
BNT—total correct responses (/60)
z
52.7 4.5 51.5 5.2 .018 53.3 4.6 54.3 4.4 .054 .003
Composite Z-score
z
22.16 1.44 21.97 1.56 .237 21.59 1.13 21.36 1.23 .234 .840
Abbreviations: BNT, Boston Naming Test; BVMT-R, Brief Visuospatial Memory Test–Revised; RL/RI-16, 16-item free/cued word learning and recall test;
SD, standard deviation; Stroop, Stroop Color-Word Interference Test; VF, Verbal Fluency.
*Intragroup Pvalue.
y
Intergroup Pvalue.
z
Specific tests used to calculate the composite Z-score for each cognitive domain.
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-10 3

visually and organoleptically indistinguishable from the
active kMCT drink, as confirmed by a blinded visual inspec-
tion and taste test. The volume of active or placebo drink to
be consumed was increased every few days from 50 mL/day
to the final dose of 250 mL/day within 2 weeks, split evenly
between two meals (usually breakfast and supper). Compli-
ance was measured by a combination of bottle count (three
to four extra bottles were provided monthly both for spillage
and as a check on total intake), daily logs, and a blinded
monthly blood test.
2.5. Neuroimaging protocol
The PET and MRI protocols were the same as those re-
ported previously [12,13,16]. On the day of the post-
intervention PET scan, 2 hours before injecting the [
11
C]-
AcAc tracer, participants consumed 125 mL of their usual
drink. PET images were acquired on a PET/computed
tomography scanner (Gemini TF, Philips Healthcare,
Eindhoven, the Netherlands). Forearm blood was
arterialized by warming with a heating pad at 44C. [
11
C]-
AcAc (370 MBq) was injected followed by a 10-minute
acquisition. One hour later, 185 MBq of [
18
F]-FDG was in-
jected with a 30-minute acquisition starting 30-minute after
injection. The image-derived input function was calibrated
against a series of blood samples as previously described
[10,11]. The [
18
F]-FDG input function was concatenated
using the image-derived input function acquired during the
scan and then cross-calibrated to plasma radioactivity (Co-
bra gamma counter, Packard).
MRIs were acquired on a 3 Tesla scanner with a 32-
channel head coil (Ingenia, Philips Healthcare, Best, the
Netherlands). For cerebral blood flow, a pseudo-continuous
Arterial Spin Labelling sequence was used: scan
duration 55 minutes and 45 seconds, two-dimensional
(2D) echo planar imaging, repetition time
54200 milliseconds, echo time 517.3 milliseconds, flip
angle 590, postlabeling delay of 2000 milliseconds, label
duration of 1650 milliseconds, 16 slices of 512 !512 of
3.0 !3.0 !4.0 mm
3
pixel size with a gap of 1 mm between
slices. A proton density image was acquired for the quanti-
fication of cerebral blood flow: repetition
time 512,000 milliseconds, echo time 517 milliseconds,
and flip angle 590. For resting state functional connectiv-
ity metrics, the protocol was: T2*-weighted 2D echo planar
imaging sequence, scan duration 55 minutes, repetition
time 52000 milliseconds, echo time 530 milliseconds,
flip angle 590, 150 averages, a 35 slices of 64 !64 of
3.0 !3.0 !3.0 mm
3
pixel size with a gap of 1 mm between
slices [38].
2.6. Image analysis
PET tracer kinetics were analyzed using PMOD 3.8 (PMOD
Technologies Ltd, Zurich, Switzerland) as previously described
[13]. The Patlak method was used to quantify the brain uptake
rate constants for both tracers (K
AcAc
,K
Glu
; minute
21
)andtheir
respective cerebral metabolic rates (CMR
AcAc
,CMR
Ketones
,
and CMR
Glu,
;mmol/100 g/minute) [12,15,39].Voxelwise
parametric images were 3D surface-projected using MIMvista
(6.4, MIM Software Inc, Cleveland, OH).
Regional and whole brain volumes and cortical thick-
nesses were determined using FreeSurfer Suite 6.0 (Marti-
nos Center for Biomedical Imaging, Cambridge, MA).
Regional volumes were normalized to the intracranial vol-
ume of each participant [40]. Cerebral blood flow was calcu-
lated using a one-compartment model (FSL version 4.1;
FMRIB, Oxford, UK) [41]. Using PMOD software, partial
volume correction was applied as described previously
[42]. Data processing of resting state functional magnetic
resonance images was performed using statistical parametric
mapping and a resting state functional MRI data analysis
toolkit [43]. Functional connectivity for the default mode
network analysis was calculated using a spherical seed
(radius 58 mm), centered at coordinates (28, 256, 26),
within the posterior cingulate cortex [38].
2.7. Laboratory methods
Plasma glucose, cholesterol, and triglycerides (Siemens
Medical Solutions USA, Inc, Deerfield, IL) as well as plasma
ketones collected during the PET scan and at monthly follow-
up [44] were analyzed by automated colorimetric assay on a
clinical chemistry analyzer (Dimension Xpand Plus;
Siemens, Deerfield, IL). Plasma BHB and AcAc were
analyzed as previously described [45,46].Plasmamedium
chain fatty acid analysis from monthly follow-up samples
was performed by ultrahigh performance liquid chromatog-
raphy (Nexera X2, Shimadzu) and tandem mass spectrometry
(API-3000, ABSciex) as previously described [47]. The other
blood metabolites were assayed at the biochemistry core lab-
oratory of CHUS. APOE genotyping was performed by real-
time polymerase chain reaction [48].
2.8. Statistics
Sample size was based on the primary outcome variable,
which was change in CMR
AcAc
and was calculated to detect
an effect size of 0.5, with an alpha risk of 5% and 90% power
(G*Power 3.1.9.2) [49]. As established from our previous
work [11,15], the coprimary outcomes—plasma ketones
and CMR
AcAc
—typically increase 2- to 3-fold on 30 g/day
of MCT, which is equivalent to an effect size of 0.5. Antic-
ipating a priori a 30% dropout during the 6-month interven-
tion, the total sample size for the study was 34 (N 517
completers per group).
Data are presented as the mean 6standard deviation. All
statistical analyses were performed using SPSS 24.0 soft-
ware (SPSS Inc, Chicago, IL). Because assumptions of ho-
mogeneity and normality of the variance were not fulfilled
for most of the dependent variables, nonparametric tests
were used. AWilcoxon signed rank test was used to compare
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-104

intragroup differences (PRE vs. POST) and a Mann-
Whitney Utest for the intergroup differences (placebo vs.
active). Linear regression was used to identify a causal
link in which variable “X” predicts the outcome variable
“Y”; for example, whether a difference in plasma or brain
ketones was associated with a change in cognitive scores.
3. Results
Of the 52 enrolled participants, n58 dropped out of the
active group (n56 were intolerant to the drink; n52 discon-
tinued for other reasons) whereas n55 dropped out of the
placebo group (n52 were intolerant; n53 discontinued
for other reasons). Thus, 75% completed the intervention,
n519 in the active group and n520 in the placebo arm
of the trial (Supplementary Fig. 1). All completers were
protocol-compliant, that is, consumed a mean of 90 68%
of the planned daily dose of the kMCT, as measured by return
bottle count. At baseline, the two groups were well matched
for age, gender, APOE4, cognitive score, education, blood
pressure, depression score, physical autonomy, clinical
chemistry, and plasma metabolites (Table 2). Average plasma
levels of the medium-chain fatty acids, octanoic acid (C8:0)
and decanoic acid (C10:0), measured at monthly intervals
were very significantly increased in the active group,
131 695 and 159 6135 mmol/L for C8:0 and C10:0, respec-
tively, compared with 5 68 and 4 66mmol/L for C8:0 and
C10:0, respectively, in the placebo group (P,.001 between
active and placebo for both C8:0 and C10:0). There were no
serious or severe adverse events in either group. About half
the participants reported at least one side effect (abdominal
or stomach discomfort [n514], reflux [n58], diarrhea
[n53], nausea [n52], bloating [n51], headache
[n51], and/or constipation [n51]). There were twice as
many reported gastrointestinal side effects in the active as
placebo group but most were transitory. There was no change
in body weight or any clinically significant changes in plasma
metabolites or clinical chemistry in either group.
3.1. Plasma and brain ketone and glucose metabolism
Compared with baseline, at the end of the study, plasma
AcAc and BHB were 221% and 262% higher, respectively,
in the active group but were unchanged in the placebo group
(Table 3). Global brain CMR of AcAc (CMR
AcAc
) was 211%
higher in the active group but did not change over 6 months
on placebo (P,.01; Fig. 1). Whole brain CMR
Ketones
(AcAc 1BHB) also did not change on the placebo but
was 230% higher post-treatment in the active group
(P,.01; Fig. 2). In the active group, CMR
AcAc
and
CMR
ketone
were 202% to 228% higher across the main brain
regions (Fig. 1,Table 3).
Table 2
Participant characteristics at enrollment
Parameters
Placebo (N 520) Active (N 519)
Intergroup PvalueMean SD Mean SD
Gender (M/F) 8/12 10/9 .582
APOE4 carrier/total sample (%) 8/19 (42%) 6/18 (33%) .429
Age (y) 75.4 6.6 73.8 6.3 .428
Education (y) 12.5 3.7 13.2 3.5 .687
GDS (/30)*7.6 4.6 6.3 6.2 .163
SMAF-E (/24)
y
1.5 2.0 2.1 4.0 .644
PASE (/793)
z
118.4 52.0 157.7 83.5 .134
McNair (/45)
x
19.2 4.9 20.8 8.8 .558
MMSE (/30)
{
27.1 2.1 27.7 2.2 .284
MoCA (/30)
#
22.4 2.4 23.5 3.5 .113
Blood pressure (systolic; mm Hg) 138.9 17.5 135.8 11.8 .422
Blood pressure (diastolic; mm Hg) 79.1 8.4 84.3 8.6 .081
Body mass index 25.8 4.0 28.2 4.3 .074
Plasma metabolites
Total cholesterol (mM) 4.8 1.1 5.0 1.0 .753
Triglycerides (mM) 0.9 0.4 1.2 0.4 .064
Creatinine (mM) 74.1 13.4 76.6 25.1 .707
Glucose (mM) 4.8 0.7 4.8 0.8 .893
Glycated hemoglobin (%) 5.8 0.5 5.6 0.3 .573
Thyroid stimulating hormone (mUI/L) 2.4 1.4 2.3 0.9 .661
Vitamin B12 (pmol/L) 375 175 384 179 .860
Abbreviation: SD, standard deviation.
*Geriatric depression screening scale [26].
y
Instrumental activity of daily living of the functional autonomy measurement system [27].
z
Physical Activity Scale for the Elderly [50].
x
McNair Frequency of Forgetting Questionnaire [51].
{
Mini-Mental State Examination [29].
#
Montreal Cognitive Assessment [30].
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-10 5

Postintervention, K
AcAc
was 9% higher in the placebo
group but unchanged in the active group (P5.028 and
.133, respectively; Supplementary Table 1). Both globally
and regionally, K
AcAc
was different between active and pla-
cebo in almost all brain regions measured.
There was no change in whole or regional brain CMR
Glu
or K
Glu
in either group (Fig. 2 and Supplementary Table 2;
all P.107). Net global brain energy uptake
(CMR
Glu
1CMR
Ketones
combined) increased from 28.2 to
29.3 mmol/100 g/minute (13.6%) in the active group
(P,.001), but did not change in the placebo group (P.50).
3.2. Cognitive outcomes
Cognitive outcomes are presented as raw scores
(Table 1). The two groups had equivalent baseline perfor-
mance on the cognitive tests. General cognition based on
MMSE and MoCA scores did not change after 6 months
in either group (data not shown; all P..1). For the
episodic memory domain (a key indicator of risk of
Table 3
Plasma ketone concentrations (mM) and regional changes in brain total ketone (acetoacetate 1beta-hydroxybutyrate combined) uptake (CMR
Ketones
;mmol/
100 g/minute) before (PRE) and at the end (POST) of the intervention
Parameters
Placebo (N 520) Active (N 519) DPlacebo versus Dactive
PRE POST Intragroup
Pvalue
PRE POST Intragroup
Pvalue
Intergroup
PvalueMean SD Mean SD Mean SD Mean SD
Plasma concentrations (mM)
Acetoacetate 132 74 112 53 .546 123 56 272 141 .001 .001
Beta-Hydroxybutyrate 215 110 180 87 .421 207 133 543 321 .001 .001
CMR
Ketones
(mmol/100 g/min)*
Frontal lobe 1.12 0.62 1.06 0.54 .970 1.14 0.77 2.59 1.75 ,.001 ,.001
Parietal lobe 1.17 0.64 1.11 0.56 .970 1.14 0.73 2.64 1.70 ,.001 ,.001
Temporal lobe 1.04 0.56 0.99 0.51 .970 1.02 0.66 2.39 1.58 ,.001 ,.001
Occipital lobe 1.24 0.71 1.17 0.63 .940 1.20 0.76 2.81 1.85 ,.001 ,.001
Cingulate cortex 0.90 0.50 0.87 0.45 .911 0.93 0.64 2.10 1.42 ,.001 ,.001
Subcortical regions 0.65 0.36 0.64 0.34 .852 0.68 0.47 1.52 0.92 ,.001 ,.001
Cortex (overall mean) 1.09 0.60 1.03 0.53 .940 1.08 0.71 2.49 1.66 ,.001 ,.001
Abbreviations: CMR, cerebral metabolic rate; SD, standard deviation.
*Calculated from brain acetoacetate uptake according to Blomqvist et al. [39] and Castellano et al. [12].
Fig. 1. Surface maps of brain AcAc uptake (CMR
AcAc
[mmol/100 g/minute]
and the rate constant of brain AcAc uptake (K
AcAc
[minute
21
] before (PRE)
and at the end of (POST) the intervention with the active or placebo drinks.
Abbreviations: AcAc, acetoacetate; CMR, cerebral metabolic rate.
Fig. 2. Whole brain glucose (CMR
Glu
) and ketone (CMR
Ketones
)metabolism
in the placebo and active groups before (PRE) and at the end of (POST) the
interventions. Whole brain CMR
Ketones
increased by 1130% at the end of the
activetreatment (P,.001) withno change in the placebo group. There was no
change in whole brain CMR
Glu
in the active or placebo group (P.687).
Mann-Whitney Utest (*P,.05). * Represents the difference between
PRE and POST in the active group. Abbreviations: AcAc, acetoacetate;
CMR, cerebral metabolic rate.
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-106

progression to AD), the active group had 20% more words
recalled on the first free recall trial of the 16-item free/cued
word learning and recall test compare to baseline (French
version, RLRI-16 test) (Z-score 511.1; P5.013),
whereas the placebo group had 9% fewer words recalled
on the delayed total recall test (Z-score 520.6;
P5.026). There was also a significant increase in the
score in first trial of the Brief Visual Memory Test-
Revised for the active group (54% higher score;
P5.027), and a tendency for improvement in preinterven-
tion to postintervention for the placebo group (P5.051).
The placebo group had a 7% lower postintervention score
on the VF categories test (P5.047). In the active group, the
number of self-corrected or noncorrected errors on the
Stroop test decreased by 44% (P5.046) and 58%
(P5.036), respectively, suggesting a modest improvement
of inhibitory capacity post-treatment. Processing speed
was 15% slower on the visual scanning task of the Trail Mak-
ing Test for the placebo group (P5.022), but remained un-
changed in the active group. Visual selective attention was
better on the number sequencing condition for the active
group post-treatment (20% less time to complete;
P5.043). After normalization, the same cognitive tests
remained statistically significant.
Comparing the placebo versus active groups, only lan-
guage as measured by the BNT showed a significant inter-
group effect for the total correct response (active
11.0 62.2 and placebo 21.3 62.0; P5.003).
3.3. Correlation of cognitive outcomes to ketone status
Scores on the following tests—Visual Scan task of Trail
Making, BNT, and VF (categories)—all improved signifi-
cantly and in direct relation to the increase in plasma ketones
post-intervention (Fig. 3; all P.043). Composite Z-scores
of processing speed (Supplemental Table 3;P5.035) and
performance on the Visual Scan task of the Trail Making
test (P5.034; not shown) also improved in direct relation
to the increase in brain ketone uptake. Both groups (active
and placebo) were included in the scatter plots (Fig. 3)
because our goal was to have enough statistical power to
determine whether there was a link between ketone concen-
tration and cognitive performance. We repeated the analysis
of each group separately and the same significant relation-
ship was observed for the active group but not for the
placebo group (data not reported).
3.4. Exploratory outcomes
Postintervention, there was no change in global brain vol-
ume or in the volume of any brain region in either group with
the exception of a 2% increase in the volume of the lateral
ventricles in the placebo group (35.8 to 36.6 mL;
P,.007). Global and regional cortical thickness, functional
connectivity, and cerebral blood flow did not change glob-
ally or regionally in either group (data not shown).
4. Discussion
The BENEFIC trial demonstrated that a kMCT drink im-
proves net brain energy status in MCI. The improvement in
brain energy status was specifically due to brain ketone up-
take doubling on the kMCT drink because the interventions
did not change brain glucose uptake in either group. Cogni-
tive scores in several domains linked to the risk of progress-
ing to AD improved significantly and in direct relation to the
increase in plasma ketones and/or brain ketone uptake on
kMCT, suggesting that these functional improvements
were because of brain energy rescue with ketones [47].
With 75% of enrolled participants completing this 6-month
intervention trial, we demonstrate that this form of long-
term ketogenic intervention is safe and feasible in MCI.
Given that brain ketone uptake is also normal in mild to
moderate AD [11,13,52,53], a kMCT drink could
potentially also have beneficial effect on cognition in AD
as previously reported [18].
The present results corroborate previous studies using
different types of ketogenic interventions in both MCI
Fig. 3. Scatter plots of the change in plasma ketones (acetoacetate 1beta-hydroxybutyrate; mM) and change in the score for the Trail Making (Visual Scan:
r520.351, P5.031), Verbal Fluency (categorical: r510.330, P5.043), and Boston Naming (r510.331, P5.042) tests. Placebo () and active (-).
N519 per group. Linear regression (P,.05).
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-10 7

[17] and mild-moderate AD [18,19]. Hence, not only is
the brain energy deficit in MCI specific to glucose but
at least partially correcting this deficit with ketones
results in cognitive improvements. A very high fat
ketogenic diet produces ketones from long-chain fatty
acids in the diet or liberated from adipose tissue, whereas
it is mostly the medium-chain fatty acid, caprylic acid,
that is ketogenic in kMCT. Hence, it is likely that it is
brain energy rescue by ketones themselves that drives
the cognitive improvement observed here rather than
any specific dietary fatty acid that changed brain mem-
brane composition. Unexpectedly, the brain’s capacity
to extract ketones from the blood (K
AcAc
) increased by
9% on the placebo but not on the active treatment. The
reason for the change in K
AcAc
in the placebo group
remains unclear at this time.
To our knowledge, only one other study has reported
the effect of kMCT versus placebo in MCI [54].That
feasibility study used a dose of 56 g/day in n52per
group who completed the intervention. The modest but
significant cognitive benefit observed in the present MCI
study is also consistent with three previous assessments
of kMCT in AD, one of which was acute and uncontrolled
[55], and the two other were placebo-controlled and of 2 to
3 months duration [18,21]. The present study had a better
matched placebo than reported by Henderson et al. [18].
We also used a dose of kMCT that came closer to closing
the brain energy gap caused by lower brain glucose up-
take; participants in the other AD studies [18,21] took a
single 20 g dose/day of kMCT whereas ours took two
15 g doses/day.
Plasma ketone half-life is relatively short, so peak
brainketoneuptakeafteradoseofkMCTis2-to
3-fold higher than brain ketone uptake when averaged
over 24 h [45]. At peak plasma ketones, the 30 g daily
dose of kMCT left approximately a 1% brain energy
deficit compared with cognitively normal older people
[13], a deficit that would be at least 3% when plasma ke-
tones are averaged over 24 h. A higher daily dose of
kMCT than 30 g, possibly up to 45 or 60 g, and/or a
formulation that improves the ketogenic potential of
MCT, is therefore probably needed to fully compensate
for the brain glucose deficit in MCI.
The BENEFIC trial had several limitations. The sample
size was not sufficient to obtain definitive cognitive results.
Despite the fact that intergroup changes in cognition were
limited to statistically better performance in the language
domain (BNT) and that this change may not be clinically sig-
nificant, the significant improvement PRE versus POST in
the kMCT group encourages us to continue recruitment to
double enrollment with the aim of better defining the impact
of kMCT on cognitive function in MCI. Also, possible
changes in measures of daily living should be measured in
a future study. Unlike the reported detrimental effect of
APOE4 status on both the ketogenic effect and cognitive
benefit of kMCT in MCI or AD [55,56], we did not see
any significant effect of APOE4 status on ketosis or
cognitive outcomes; however, the present study was not
adequately powered to assess this.
Most participants reported some gastrointestinal side ef-
fects during the project, especially in the active group.
Providing the MCT as an emulsion and recommending it
be taken with meals reduced the frequency and severity of
these side effects but eight participants still dropped out prin-
cipally for this reason. Tolerance and convenience should be
taken into account when formulating a new approach with a
higher dose of kMCT.
In conclusion, we demonstrate here for the first time
that a dose of 30 g/day of kMCT taken for 6 months pro-
vides enough ketones to significantly improve brain en-
ergy status in MCI. Several aspects of cognitive
functionalsoimprovedindirectrelationtotheincrease
inbrainenergystatusachievedwiththekMCT.Itre-
mainstobeseenwhetheralargersamplesizewill
confirm the cognitive benefit of this dose of kMCT or
whether a higher dose is required. Nevertheless, we
show here that long-term clinical trials with kMCT and
energy-equivalent placebo are feasible in older people.
Further research to delay aging-related cognitive decline
by optimizing brain energy rescue with ketones is war-
ranted.
Acknowledgments
The authors thank Dr S
ebastien Tremblay, Christine
Brodeur-Dubreuil, Marie Christine Morin, Louise-Andr
ee
Lambert, Odette Baril, Audrey Perreault,

Eric Lavall
ee,
and the clinical team at the Sherbrooke Molecular Imaging
Center for technical assistance.
This project was funded by the Part-the-Cloud program of
the Alzheimer Association USA, MITACS, and the Uni-
versit
e de Sherbrooke (University Research Chair to SCC).
Author contributions: M.F., C.A.C., and S.C.C. conceived
the study design for the BENEFIC trial. M.F., C.A.C.,
and F.L. ran the cognitive assessments and analyzed and in-
terpreted the cognitive data. V.S.P., C.V., M.B., M.D.,
K.W., M.R., and M.L. contributed to experimental method-
ology and image and biological analyses. C.A.C. conduct-
ed the statistical analyses. E.C., C.A.C., M.F., and V.S.P.
conducted the PET and MRI scans. C.B., T.F., and E.T. pro-
vided medical supervision and assessments throughout the
study. S.C.C. drafted and revised the manuscript. All coau-
thors reviewed and commented on the manuscript before
submission.
Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.jalz.2018.12.017.
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-108

RESEARCH IN CONTEXT
1. Systematic review: All peer-reviewed articles avail-
able on PubMed on the subject of ketones or brain
ketone uptake and Alzheimer’s, mild cognitive
impairment (MCI) or dementia were reviewed. We
found no published work describing measurement of
the relationship between brain ketone uptake and
cognitive outcomes after a ketogenic intervention in
MCI.
2. Interpretation: This randomized controlled trial
demonstrated for the first time that a ketogenic me-
dium chain triglyceride drink increased certain
cognitive outcomes in MCI in direct relation to the
net change in brain energy status. Our results support
previous reports showing that various ketogenic in-
terventions can improve cognitive outcomes in both
MCI and Alzheimer’s disease.
3. Future directions: This study was powered to assess
the change in brain energy status. A larger sample
size is required to determine the robustness of the
cognitive improvement we observed. The dose of
ketogenic medium chain triglyceride needed to opti-
mize cognitive outcomes may need to be higher.
References
[1] Cunnane SC, Courchesne-Loyer A, St-Pierre V, Vandenberghe C,
Pierotti T, Fortier M, et al. Can ketones compensate for deteriorating
brain glucose uptake during aging? Implications for the risk and treat-
ment of Alzheimer’s disease. Ann N Y Acad Sci 2016;1367:12–20.
[2] Veech RL, Chance B, Kashiwaya Y, Lardy HA, Cahill GF Jr. Ketone
bodies, potential therapeutic uses. IUBMB Life 2001;51:241–7.
[3] Cunnane SC, Courchesne-Loyer A, Vandenberghe C, St-Pierre V,
Fortier M, Hennebelle M, et al. Can ketones help rescue brain fuel sup-
ply in later life? Implications for cognitive health during aging and the
treatment of Alzheimer’s disease. Front Mol Neurosci 2016;9:1–21.
[4] Henderson ST. Targeting diminished cerebral glucose metabolism for
Alzheimer’s disease. Drug Discov World 2013;14:16–20.
[5] Mosconi L, Brys M, Switalski R, Mistur R, Glodzik L, Pirraglia E, et al.
Maternal family history of Alzheimer’s disease predisposes to reduced
brain glucose metabolism. Proc Natl Acad Sci U S A 2007;
104:19067–72.
[6] Swerdlow RH, Burns JM, Khan SM. The Alzheimer’s disease mito-
chondrial cascade hypothesis: progress and perspectives. Biochim Bio-
phys Acta 2014;1842:1219–31.
[7] Maalouf M, Rho JM, Mattson MP. The neuroprotective properties of
calorie restriction, the ketogenic diet, and ketone bodies. Brain Res
Rev 2009;59:293–315.
[8] Owen OE, Morgan AP, Kemp HG, Sullivan JM, Herrera MG,
Cahill GF Jr. Brain metabolism during fasting. J Clin Invest 1967;
46:1589–95.
[9] Drenick EJ, Alvarez LC, Tamasi GC, Brickman AS. Resistance to
symptomatic insulin reactions after fasting. J Clin Invest 1972;
51:2757–62.
[10] Nugent S, Tremblay S, Chen KW, Ayutyanont N, Roontiva A,
Castellano CA, et al. Brain glucose and acetoacetate metabolism: a
comparison of young and older adults. Neurobiol Aging 2014;
35:1386–95.
[11] Castellano CA, Nugent S, Paquet N, Tremblay S, Bocti C, Lacombe G,
et al. Lower brain 18F-fluorodeoxyglucose uptake but normal
11C-acetoacetate metabolism in mild Alzheimer’s disease dementia.
J Alzheimers Dis 2015;43:1343–53.
[12] Castellano CA, Paquet N, Dionne IJ, Imbeault H, Langlois F,
Croteau E, et al. A 3-month aerobic training program improves
brain energy metabolism in mild Alzheimer’s disease: preliminary
results from a neuroimaging study. J Alzheimers Dis 2017;
56:1459–68.
[13] Croteau E, Castellano CA, Fortier M, Bocti C, Fulop T, Paquet N, et al.
A cross-sectional comparison of brain glucose and ketone metabolism
in cognitively healthy older adults, mild cognitive impairment and
early Alzheimer’s disease. Exp Gerontol 2018;107:18–26.
[14] Hasselbalch SG, Madsen PL, Hageman LP, Olsen KS, Justesen N,
Holm S, et al. Changes in cerebral blood flow and carbohydrate meta-
bolism during acute hyperketonemia. Am J Physiol 1996;
270:E746–51.
[15] Courchesne-Loyer A, Croteau E, Castellano CA, St-Pierre V,
Hennebelle M, Cunnane SC. Inverse relationship between brain
glucose and ketone metabolism in adults during short-term moderate
dietary ketosis: a dual tracer quantitative positron emission tomogra-
phy study. J Cereb Blood Flow Metab 2017;37:2485–93.
[16] Croteau E, Castellano C-A, Richard MA, Fortier M, Nugent S,
Lepage M, et al. Ketogenic medium chain triglycerides increase brain
energy metabolism in Alzheimer’s disease. J Alzheimers Dis 2018;
64:551–61.
[17] Krikorian R, Shidler MD, Dangelo K, Couch SC, Benoit SC, Clegg DJ.
Dietary ketosis enhances memory in mild cognitive impairment. Neu-
robiol Aging 2012;33:425.e19–425.e27.
[18] Henderson ST, Vogel JL, Barr LJ, Garvin F, Jones JJ, Costantini LC.
Study of the ketogenic agent AC-1202 in mild to moderate Alz-
heimer’s disease: A randomized, double-blind, placebo-controlled,
multicenter trial. Nutr Metab (Lond) 2009;6:31.
[19] Taylor MK, Sullivan DK, Mahnken JD, Burns JM, Swerdlow RH.
Feasibility and efficacy data from a ketogenic diet intervention in
Alzheimer’s disease. Alzheimers Dement 2018;4:28–36.
[20] Craft S, Neth BJ, Mintz A, Sai K, Shively N, Dahl D, et al. Ketogenic
diet effects on brain ketone metabolism and Alzheimer’s disease CSF
biomarkers. Alzheimers Dement 2016;12:P342–3.
[21] Ota M, Matsuo J, Ishida I, Takano H, Yokoi Y, Hori H, et al. Effects of a
medium-chain triglyceride-based ketogenic formula on cognitive
function in patients with mild-to-moderate Alzheimer’s disease.
Neurosci Lett 2019;690:232–6.
[22] Bach AC, Babayan VK. Medium-chain triglycerides: An update. Am J
Clin Nutr 1982;36:950–62.
[23] Martens B, Pfeuffer M, Schrezenmeir J. Medium-chain triglycerides.
Int Dairy J 2006;16:1374–82.
[24] Vandenberghe C, St-Pierre V, Pierotti T, Fortier M, Castellano C-A,
Cunnane SC. Tricaprylin alone increases plasma ketone response
more than coconut oil or other medium chain triglycerides: An acute
crossover study in healthy adults. Curr Dev Nutr 2017;1:e000257.
[25] Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern
Med 2004;256:183–94.
[26] Yesavage JA, Brink TL, Rose TL. Development and validation of a
geriatric depression screening scale: a preliminary report. J Psychiatr
Res 1982;17:37–49.
[27] H
ebert R, Guilbault J, Desrosiers J, Dubuc N. The functional auton-
omy measurement system (SMAF): a clinical-based instrument for
measuring disabilities and handicaps in older people. Can Geriatr J
2001;4:141–7.
[28] American Psychiatric Association. In: Diagnostic and Statistical
Manual of Mental Disorders. 5th ed. Washington, DC: American Psy-
chiatric Association; 2013 (DSM-5).
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-10 9

[29] Folstein MF, Folstein SE, McHugh PR. ‘Mini mental state’. A prac-
tical method for grading the cognitive state of patients for the clinician.
J Psychiatr Res 1975;12:189–98.
[30] Nasreddine ZS, Phillips NA, B
edirian V, Charbonneau S,
Whitehead V, Collin I, et al. The Montreal Cognitive Assessment,
MoCA: a brief screening tool for mild cognitive impairment. J Am
Geriatr Soc 2005;53:695–9.
[31] Van der Linden M, Coyette F, Poitrenaud JGREMEM. L’
epreuve de
Rappel Libre/Rappel Indic
e
a 16 items (RL/RI-16). In: Van der
Linden M, Agniel A, eds. L’
evaluation des troubles de la m
emoire

episodique (avec leur 
etalonnage). Marseille (France): Solal; 2004.
p. 25–47.
[32] Benedict R. Brief Visuospatial Memory Test-Revised professional
manual 1997. Odessa, FL: Psychological Assessment Resources; 1997.
[33] Delis D, Kaplan E, Kramer J. In: Delis-Kaplan Executive Function
System (D-KEFS). San Antonio, TX (USA): The Psychological Cor-
poration; 2001.
[34] Wechsler D. In: Wechsler Memory Scale. 3rd ed. San Antonio, TX:
The Psychological Corporation; 1997.
[35] Kaplan E, Goodglass H, Weintraub S. The Boston Naming Test 1983.
Philadelphia: Lea & Febiger; 1983.
[36] Zec RF, Burkett NR, Markwell SJ, Larsen DL. Normative data strati-
fied for age, education, and gender on the Boston Naming Test. Clin
Neuropsychol 2007;21:617–37.
[37] Dion M, Potvin O, Belleville S, Ferland G, Renaud M, Bherer L, et al.
Normative data for the Rappel libre/Rappel indice a 16 items (16-item
Free and Cued Recall) in the elderly Quebec-French population. Clin
Neuropsychol 2015;28(Suppl 1):S1–19.
[38] Bernier M, Croteau E, Castellano CA, Cunnane SC, Whittingstall K.
Spatial distribution of resting-state BOLD regional homogeneity as a
predictor of brain glucose uptake: A study in healthy aging. Neuro-
image 2017;150:14–22.
[39] Blomqvist G, Thorell JO, Ingvar M, Grill V, Widen L, Stone-
Elander S. Use of R-beta-[1-11C]hydroxybutyrate in PET studies of
regional cerebral uptake of ketone bodies in humans. Am J Physiol
1995;269:E948–59.
[40] Whitwell JL, Crum WR, Watt HC, Fox NC. Normalization of cere-
bral volumes by use of intracranial volume: Implications for longi-
tudinal quantitative MR imaging. AJNR Am J Neuroradiol 2001;
22:1483–9.
[41] Alsop DC, Detre JA, Golay X, Gunther M, Hendrikse J, Hernandez-
Garcia L, et al. Recommended implementation of arterial spin-
labeled perfusion MRI for clinical applications: A consensus of the
ISMRM perfusion study group and the European consortium for
ASL in dementia. Magn Reson Med 2015;73:102–16.
[42] Muller-Gartner HW, Links JM, Prince JL, Bryan RN, McVeigh E,
Leal JP, et al. Measurement of radiotracer concentration in brain
gray matter using positron emission tomography: MRI-based correc-
tion for partial volume effects. J Cereb Blood Flow Metab 1992;
12:571–83.
[43] Song XW, Dong ZY, Long XY, Li SF, Zuo XN, Zhu CZ, et al. REST: A
toolkit for resting-state functional magnetic resonance imaging data
processing. PLoS One 2011;6:e25031.
[44] Courchesne-Loyer A, Fortier M, Tremblay-Mercier J, Chouinard-
Watkins R, Roy M, Nugent S, et al. Stimulation of mild, sustained
ketonemia by medium-chain triacylglycerols in healthy humans: esti-
mated potential contribution to brain energy metabolism. Nutrition
2013;29:635–40.
[45] St-Pierre V, Courchesne-Loyer A, Vandenberghe C, Hennebelle M,
Castellano CA, Cunnane SC. Butyrate is more ketogenic than leucine
or octanoate-monoacylglycerol in healthy adult humans. J Funct Foods
2017;32:170–5.
[46] Courchesne-Loyer A, Lowry C-M, St-Pierre V, Vandenberghe C,
Fortier M, Castellano C-A, et al. Emulsification increases the acute
ketogenic effect and bioavailability of medium-chain triglycerides in
humans protein, carbohydrate, and fat metabolism. Curr Dev Nutr
2017;1:e000851.
[47] Morris AA. Cerebral ketone body metabolism. J Inherit Metab Dis
2005;28:109–21.
[48] Koch W, Ehrenhaft A, Griesser K, Pfeufer A, Muller J, Schomig A,
et al. TaqMan systems for genotyping of disease-related polymor-
phisms present in the gene encoding apolipoprotein E. Clin Chem
Lab Med 2002;40:1123–31.
[49] Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible sta-
tistical power analysis program for the social, behavioral, and biomed-
ical sciences. Behav Res Methods 2007;39:175–91.
[50] Washburn RA, Smith KW, Jette AM, Janney CA. The Physical Activ-
ity Scale for the Elderly (PASE): development and evaluation. J Clin
Epidemiol 1993;46:153–62.
[51] McNair D, Kahn R. Self-assessment of cognitive deficits. In: Crook T,
Fems S, Bartus R, eds. Assessment in geriatric psychopharmacology.
New Canaan, CT: Mark Powley Associates; 1983. p. 137–43.
[52] Lying-Tunell U, Lindblad BS, Malmlund HO, Persson B. Cerebral
blood flow and metabolic rate of oxygen, glucose, lactate, pyruvate,
ketone bodies and amino acids. II. Presenile dementia and normal-
pressure hydrocephalus. Acta Neurol Scand 1981;63:337–50.
[53] Ogawa M, Fukuyama H, Ouchi Y, Yamauchi H, Kimura J. Altered energy
metabolism in Alzheimer’s disease. J Neurol Sci 1996;139:78–82.
[54] Rebello CJ, Keller JN, Liu AG, Johnson WD, Greenway FL. Pilot
feasibility and safety study examining the effect of medium chain tri-
glyceride supplementation in subjects with mild cognitive impairment:
a randomized controlled trial. BBA Clin 2015;3:123–5.
[55] Reger MA, Henderson ST, Hale C, Cholerton B, Baker LD,
Watson GS, et al. Effects of
b
-hydroxybutyrate on cognition in
memory-impaired adults. Neurobiol Aging 2004;25:311–4.
[56] Henderson ST, Poirier J. Pharmacogenetic analysis of the effects of
polymorphisms in APOE, IDE and IL1B on a ketone body based ther-
apeutic on cognition in mild to moderate Alzheimer’s disease; a ran-
domized, double-blind, placebo-controlled study. BMC Med Genet
2011;12:137.
M. Fortier et al. / Alzheimer’s & Dementia -(2019) 1-1010