Available online at www.sciencedirect.

com

ScienceDirect
Journal of Nutritional Biochemistry 33 (2016) 91 102

Effect of dietary docosahexaenoic acid (DHA) in phospholipids or triglycerides on

brain DHA uptake and accretion

Alex P. Kitson a , Adam H. Metherel a , Chuck T. Chen a, 1 , Anthony F. Domenichiello a , Marc-Olivier Trpanier a ,
Alvin Berger b, c , Richard P. Bazinet a,
a
Department of Nutritional Sciences, University of Toronto, Toronto, Ontario, M5S3E2, Canada
b
Arctic Nutrition AS, NO-6155, rsta, Norway
c
Department of Food Science & Nutrition, University of Minnesota, St. Paul, MN, 55108-1038, USA

Received 6 October 2015; received in revised form 27 January 2016; accepted 11 February 2016

Abstract

Tracer studies suggest that phospholipid DHA (PL-DHA) more effectively targets the brain than triglyceride DHA (TAG-DHA), although the mechanism and
whether this translates into higher brain DHA concentrations are not clear. Rats were gavaged with [U-3H]PL-DHA and [U-3H]TAG-DHA and blood sampled over
6 h prior to collection of brain regions and other tissues. In another experiment, rats were supplemented for 4 weeks with TAG-DHA (fish oil), PL-DHA (roe PL)
or a mixture of both for comparison to a low-omega-3 diet. Brain regions and other tissues were collected, and blood was sampled weekly. DHA accretion rates
were estimated using the balance method. [U-3H]PL-DHA rats had higher radioactivity in cerebellum, hippocampus and remainder of brain, with no differences
in other tissues despite higher serum lipid radioactivity in [U-3H]TAG-DHA rats. TAG-DHA, PL-DHA or a mixture were equally effective at increasing brain DHA.
There were no differences between DHA-supplemented groups in brain region, whole-body, or tissue DHA accretion rates except heart and serum TAG where the
PL-DHA/TAG-DHA blend was higher than TAG-DHA. Apparent DHA -oxidation was not different between DHA-supplemented groups. This indicates that more
labeled DHA enters the brain when consumed as PL; however, this may not translate into higher brain DHA concentrations.
2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Docosahexaenoic acid; Brain; Triacylglycerol; Phospholipid; Serum; Uptake

1. Introduction Alzheimer's disease [711] (reviewed in [12]), major depression

Docosahexaenoic acid (DHA, 22:6n-3) is highly concentrated in the
brain where it composes approximately 10% of all fatty acids [1]. DHA
regulates many neurological processes including neuronal survival [2],
synaptogenesis [3], neuronal plasticity [4] and neuroinammation [5].
Experimental depletion of brain DHA in animal studies is associated
with behavioral impairments [reviewed in [6]], and some neurode-
generative and neuropsychiatric disorders have been associated with
low brain DHA content. For example, lower postmortem brain DHA in
select brain regions and lipid classes has been found in patients with
This work was funded by Arctic Nutrition AS, NO-6155 rsta, Norway.
Funding source was involved in study design, but not in analysis, data
handling or manuscript preparation. Alex P. Kitson received a fellowship from
the Canadian Institutes of Health Research. Richard P. Bazinet holds a Canada
Research Chair in Brain Lipid Metabolism.
Corresponding author at: Department of Nutritional Sciences, University
of Toronto, Fitzgerald Building, 150 College St. Room 306, Toronto, Ontario,
M5S 3E2, Canada. Tel.: + 1-416-946-8276.
E-mail address: richard.bazinet@utoronto.ca (R.P. Bazinet).
1
Present Address: National Institute on Alcohol Abuse and Alcoholism,
National Institutes of Health, Bethesda, MD, 20,8929410, USA.
[13,14], bipolar disorder [15] and schizophrenia [16], although some
studies show no differences [1,1719]. This has led the eld to
speculate that strategies to increase brain DHA concentration may be
effective in treatment/prevention of neurodegenerative/neuropsychi-
atric disorders, assuming that low brain DHA is causally linked to the
disease progression [2023].
The brain can synthesize DHA from -linolenic acid (ALA, 18:3n-3)
[24,25]; however, the synthesis rate (1.5 nmol/g/d) [24] is much
lower than the total rate of brain DHA accretion (190 nmol/g/d) [26] in
rats, indicating that uptake from serum DHA is likely the main source
of DHA for the brain [24]. The plasma lipid pool from which the brain
takes up DHA is not agreed upon; however, there is evidence that
plasma unesteried DHA is the major contributor [25,27,28], although
other pools may contribute, such as lysophosphatidylcholine [29,30].
Dietary interventions that target serum lipid pools important for brain
DHA uptake may be effective in targeting brain DHA.
Dietary DHA is typically esteried to TAG, although supplements
with a high proportion of DHA esteried to phospholipids (PL), such as
krill oils and sh roe phospholipid extracts, are increasingly common.
Several studies have attempted to compare the bioavailability of TAG-
or PL-enriched sources of DHA based on responses of blood
biomarkers and have concluded either higher bioavailability of PL-

http://dx.doi.org/10.1016/j.jnutbio.2016.02.009
0955-2863/ 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
92 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102
DHA [3133] or equal efcacy [3437], although this area is
controversial [3841].
There is also some evidence that dietary PL-DHA may target the brain
more effectively than TAG-DHA. Liu et al. orally administered [U-13C]DHA
esteried in the sn-2 position of TAG or PC to neonatal piglets and found
approximately twofold greater labeled DHA in gray matter and
synaptosomes [42] of [U-13C]DHA-PC gavaged piglets. Similarly, Graf
et al. orally administered [1-14C]DHA esteried to sn-2 of PC and TAG and
found higher radioactivity in brain regions of 10-week-old rats gavaged
with [1-14C]DHA-PC, but not in 2- or 4-week-old rats [43].
Taken together, it appears that dietary PL-DHA may more
effectively target the brain relative to TAG-DHA; however, it is not
known if this translates into higher brain DHA concentration following
supplementation with a source of PL-DHA compared with TAG-DHA. It
is possible that ingested PL-DHA may more effectively enrich serum
lipid pools that target brain DHA compared with TAG-DHA, as there is
some evidence that dietary PL and TAG target different serum lipid
pools upon digestion and absorption [44,45]. However, a detailed
time-course comparing the serum lipid class enrichment following
ingestion of DHA-containing PL and TAG tracers has not been
performed, nor has it been assessed whether acute postprandial
serum lipid class enrichment kinetics translate into differences in
serum lipid class DHA concentration in response to semichronic PL- or
TAG-DHA supplementation.
Therefore, the goal of this study was to examine whether the
higher brain tracer uptake of ingested PL-DHA compared to TAG-DHA
translated into higher brain DHA content following 4 weeks of
supplementation with a source of PL-DHA, TAG-DHA or a mixture of
the two. In addition, differences in serum lipid pool enrichment
following ingestion of PL-DHA and TAG-DHA tracer were character-
ized and compared to changes in serum lipid pool DHA concentration
following dietary PL-DHA vs. TAG-DHA supplementation. Differences
in other tissues were also assessed.
2. Material and methods
2.1. Radiotracer gavage study tracers
Sn-1,3-dpalmitoyl-2-docosahexaenoylglycerol (TAG-DHA) and 1-palmitoyl-2-doc-
osahexaenoyl-sn-glycero-3-phosphocholine (DHA-PC) were purchased from Avanti
Polar Lipids (Alabaster, AL, USA) and were uniformly labeled with 3H by Moravek
Biochemicals (Brea, CA, USA) to form [U-3H]TAG-DHA [U-3H]DHA-PC, respectively.
Radiochemical purity of the tracers was N99.8% as determined by Moravek Biochemicals
and the specic activity of tracers was 1.1 mCi/mmol for [U-3H]TAG-DHA and 1.6 mCi/
mmol for [U-3H]DHA-PC. Tracers were shipped and dissolved in toluene and, prior to
use, were dissolved in olive oil (Sigma-Aldrich, St. Louis, MO, USA) following
evaporation of toluene under a stream of N2. The tracer was dissolved to a nal
concentration of 500 Ci/ml for [U-3H]TAG-DHA and 300 Ci of [U-3H]DHA-PC.
Different tracer concentrations were due to availability of material.
2.2. Radiotracer gavage study animals and diets
All procedures performed on animals were approved by the University of Toronto
Animal Care Committee in accordance with guidelines set by the Canadian Council of
Animal Care. Animals were housed with ad libitum access to food and water with a
12:12-h light:dark cycle and controlled temperature of 221 C.
Male LongEvans rats arrived at the University of Toronto animal care facility at
10 days of age with lactating dams. Dams were immediately placed on a low omega-3
diet with 0.140.01% of fatty acids as n-3 PUFA (Dyet's Inc., Bethlehem Pennsylvania)
modied from the AIN-93G formulation [46] as shown in Table 1. ALA content of the diet
was consistent with previous work [46]. Pups were weaned onto the same diet at
21 days of age. Pups were maintained on this diet until 11 weeks of age when they
underwent carotid artery cannulation. Briey, rats were anesthetized through
isourane inhalation (5% induction, 12.5% maintenance) followed by subcutaneous
injection of ketoprofen (5 mg/kg; Merial Canada Inc., Baie d'Urf, QC, Canada), and
incision sites (scapular, superior thoracic) were shaved and disinfected. Carotid artery
was isolated by blunt dissection and cannulated with sterile polyethylene catheter (PE
50, Intramedic; Becton Dickinson, Franklin Lakes, NJ, USA) tipped with 1.5 cm of
silicone tubing (silicone tubing 0.020 in. inner diameter and 0.037 in. outer diameter,
VWR; Mississauga, ON, Canada). The cannula was externalized in the scapular region
and a skin button (Braintree Scientic, Braintree, MA, USA) was sutured to facilitate
assembly of the sampling system.
After a 24-h recovery from surgery, a 250-l baseline blood sample was taken
followed by ushing the line with 1% heparinized saline to ensure line patency
throughout the experiment. Rats were then gavaged with 500 l of one of the two
tracers (n=4 for TAG-DHA, n=5 for PC-DHA) dissolved in olive oil providing 454.5 and
192.5 mmol of the [U-3H]TAG-DHA (500 Ci) and [U-3H]DHA-PC (308 Ci), respec-
tively. The rat was then tethered to a counterbalanced sampling system (Instech
Laboratories, Plymouth Meeting, PA, USA). Blood samples (250 l) were taken 30, 60,
90, 120, 150, 180, 210, 240, 270, 300, 330 and 360 min postgavage. In the event of
difculty obtaining blood draw from the cannula, a tail-vein blood draw was performed.
Blood was centrifuged at 10,000 g for isolation of serum and immediately frozen at
80 C until analysis. Rats were euthanized by lethal injection of T-61 (embutramide
200 mg/ml, mebezodium iodide 50 mg/ml, tetrocaine hydrochloride 5 mg/ml admin-
istered at 300-l/kg body weight, Merck Animal Health, Kirkland, QUE, Canada) via the
carotid catheter 360 min postgavage, and brain regions (cortex, cerebellum, hippo-
campus, striatum, brainstem and rest of brain) as well as liver, perirenal adipose tissue,
heart, small intestine (5 cm inferior to pyloric sphincter) and abdominal skin were
collected, rinsed in saline and frozen at 80 C until analysis.
2.3. Radiotracer gavage study lipid analysis
Lipids were extracted from serum and tissues by the method of Folch, Lees and
Sloane Stanley [47] using 2:1:0.8 chloroform:methanol:0.88% KCl (v/v/v) as previously
described [48,49]. Thin-layer chromatography was used to isolate brain region PL as
well as to collect lipid classes in serum and liver (PL, monoacylglycerols, co-eluting
diacylglycerols/cholesterol, nonesteried fatty acids, TAGs, cholesteryl esters) using
silica G plates (silica gel 60, EMD Millipore, MA, USA) and a mobile phase of 60:40:2
heptane:diethyl ether:glacial acetic acid (v/v/v). Lipid classes were visualized by
spraying resolved plates with 0.1% (w/v) 8-anilino-1-naphthalenesulfonic acid and
identied in relation to mobility of a reference standard mixture containing lipid classes
of interest. Only total lipid extracts of heart, adipose, skin and intestine were analyzed.
Radioactivity of total lipid extracts and lipid fractions was measured on a Hewlett-
Packard Tri-Carb 2900TR Liquid Scintillation Analyzer with detector efciency of
45.81%. Radioactive data were normalized to percentage of gavaged dose. Unidirec-
tional incorporation coefcients were also determined [50] using the average area
under the curve for radioactivity in unesteried fatty acids, esteried fatty acids and
total serum lipid radioactivity according to the following equation
C
i T
k
i 1
0 C
si dt
T
Where C i(T) is the radioactivity of brain region i at the end of the experiment, and
T0C sdt is the area under the curve of the serum radioactivity of serum lipid i. k* was
calculated for unesteried fatty acids, esteried lipids and total lipid radioactivity.
2.4. PL vs. TAG-DHA supplementation study animals and diets
Ten-day-old male LongEvans rat pups with lactating dams arrived at the
University of Toronto animal care facility, and the dams were immediately placed on
a low n-3 PUFA diet (same as used in radiotracer study), and pups were weaned onto the
low n-3 PUFA diet at 21 days of age. At 11 weeks of age, one group of rats was sacriced
by high-energy head-focused microwave xation (13.5 kW for 1.6 s; Cober Electronics
Inc., Stratford, CT, USA [48,49,51]) following collection of tail vein blood, and brain
regions (as in radiotracer study), liver, heart, adipose, small intestine, skin and perirenal
adipose were collected to estimate baseline DHA concentrations (n=8). The rest of the
rats were randomized to one of four experimental diets: (a) a low n-3 PUFA diet
supplemented with olive oil (OO diet; n=8); (b) a diet supplemented with sh oil (FO
diet; n=8); (c) a diet supplemented with herring caviar phospholipid concentrate (PLC
diet; n=8) with 81% of DHA esteried to phospholipids; or (d) a diet supplemented
with a blend of FO and PLC (FO+PLC diet; n=7) with 30% of the DHA esteried to
phospholipids (lipid class composition analysis performed by Mylneeld, Dundee,
Scotland). Oil sources and fatty acid concentration of diets are shown in Table 1. DHA-
containing oils were prepared by Arctic Nutrition (fatty acid composition in
Supplemental Table 1). All DHA-supplemented diets (FO, PLC, FO+PLC) were
formulated such that DHA would comprise 2% of total fatty acid mass, as this has
been suggested to be the dietary DHA content at which brain DHA is maximal [52]. Rats
were maintained on experimental diets for 4 weeks (from postnatal week 11 to week
15) with body weight, 24-h food intake and serum DHA (sampled via tail vein)
measured weekly. At the conclusion of supplementation, blood was sampled from the
tail vein, and rats were killed by high-energy head-focused microwave xation, and
samples were collected identically to baseline rats. Carcasses and 24-h fecal samples
were also collected and stored at 80 C.
2.5. PL-DHA vs. TAG-DHA supplementation study lipid analysis
Lipids were extracted from serum, tissues and feces as described for the radiotracer
study. Carcasses were homogenized in a Waring industrial blender (model CB15) with a
4-L stainless steel pitcher (Hendrix Restaurant Equipment and Supplies, Brockville, ON,
USA) with water to facilitate homogenization, and lipids were extracted in duplicate.
For all samples, an internal standard was added prior to lipid extraction for
 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102 93

Table 1
Fat source and fatty acid composition of rodent diets
Name Low n-3 PUFA Olive oil Fish oil Phospholipid concentrate Fish oil + phospholipid concentrate

Fat source g/kg diet

Hydrogenated coconut oil 66.2 50.0 50.0 50.0 50.0
Safower oil 33.8 33.8 33.8 33.8 33.8
Olive oil 0 16.2 8.8 3.4 7.85
Fish oil 0 0 7.4 0 0
Phospholipid concentrate 0 0 0 12.8 0
Phospholipid concentrate + sh oil 0 0 0 0 8.35
Total 100 100 100 100 100

Fatty acid % composition 1

C 10:0 4.440.54 2.360.45 2.320.25 2.380.08 2.350.11
C 12:0 37.450.98 25.940.51 26.50.45 27.570.11 27.140.2
C 14:0 14.210.47 10.270.12 10.980.12 11.470.08 11.120.12
C 16:0 8.900.23 9.960.19 9.620.06 10.580.14 9.850.04
C 18:0 7.380.18 6.590.17 6.770.12 7.080.1 6.820.06
C 20:0 0.170.01 0.220.005 0.230.01 0.20.01 0.220.001
SFA 75.110.56 55.781.07 56.640.69 59.530.23 57.690.45
C 16:1 0.060.01 0.350.01 0.470.004 0.490.001 0.40.003
C 18:1n-7 0.190.01 0.680.03 0.580.02 0.610.01 0.570.01
C 18:1n-9 5.250.15 16.550.39 12.250.21 9.640.07 11.720.15
C 20:1n-9 0.060.0001 0.10.01 0.340.01 0.190.004 0.150.01
C 22:1n-9 0.100.01 0.090.001 0.10.005 0.090.01 0.090.004
C 24:1n-9 TR 0.030.005 0.130.005 0.090.01 0.110.003
MUFA 5.70.17 17.790.41 13.870.25 11.140.07 13.060.14
C 18:2n-6 18.90.36 26.090.67 25.870.4 25.880.2 25.760.32
C 20:4n-6 TR TR TR TR TR
C 22:4n-6 0.10.01 0.050.013 0.080.01 0.040.01 0.060.01
N-6 19.050.38 26.180.66 26.080.39 26.020.2 25.940.33
C 18:3n-3 0.080.003 0.190.01 0.180.003 0.190.002 0.170.004
C 20:5n-3 TR TR 0.750.014 0.910.02 0.710.012
C 22:5n-3 TR TR 0.230.004 0.070.001 0.170.003
C 22:6n-3 0.050.002 0.050.002 2.240.04 2.120.04 2.250.04
N-3 0.140.01 0.250.01 3.410.05 3.310.04 3.310.06
PUFA 19.190.39 26.430.66 29.490.44 29.330.2 29.250.32
N-6/N-3 135.13.35 104.310.61 7.650.05 7.860.12 7.830.19

TR: trace (b0.05% of fatty acids), SFA: saturated fatty acids, MUFA: monounsaturated fatty acids, PUFA: polyunsaturated fatty acids.
1
Data are mean SD from triplicate analysis in our laboratory. Levels of 14:1, 18:3n-6, 20:2n-6, 20:4n-6, 22:4n-6, 22:5n-6 were measured and found to be negligible (b0.05% of
fatty acids).

phospholipids (1,2-diheptadecanoyl-sn-3-phosphatidylcholine; Avanti polar lipids), Whole-body apparent DHA -oxidation rates were determined using the following calculation:
TAGs (triheptadecanoate, Nuchek Prep) unesteried fatty acids (heptadecanoic acid,
Nuchek Prep) or cholesteryl esters (cholesteryl heptadecanoate, Nuchek Prep). Neutral Apparent daily  oxidation rate
lipid classes were isolated from serum lipid extracts by TLC as described for the Dietary DHA consumed  whole body DHA accretion  fecal DHA excretion
radiotracer study. Lysophosphatidylcholine was isolated by TLC on silica H plates 3
number of days of supplementation
(Analtech, Newark, DE, USA) with a mobile phase of chloroform:methanol:acetic
acid:acetone: water (40:25:7:4:2 by vol) [53]. Fatty acid methyl esters (FAMEs) were
prepared by heating at 100 C for 1 h in 14% boron triouride in methanol [54]. 2.8. Statistics

2.6. PL-DHA vs. TAG-DHA supplementation study gas chromatography Statistical analysis was performed using SPSS Statistics Version 22 and Prism Version
4. Signicance was inferred at Pb.05. Comparison of tissue radioactivity between
FAMEs were resolved on a Varian 430 gas chromatograph- FID (Varian, Lake Forest, [U- H]TAG-DHA and [U-3H]PL-DHA was done by independent samples t test. The effects of
3

CA, USA) with a DB-23 capillary column (30 m0.25 mm interior diameter 0.25-m
lm thickness; Agilent Technologies, Mississauga, ON, USA) with 1-ul injection volume
running in splitless mode. Gas chromatography conditions were as previously described
[48,55]. Briey, injector and detector ports were 250 C, temperature program was
50 C for 2 min, increasing at 20 C/min, and held at 170 C for 1 min, then increased at
3 C/min and held at 212 C for 5 min to complete the run at 32 min with helium as
carrier gas with ow rate of 0.7 ml/min. Peaks were identied in relation to a reference
standard mixture (GLC 455 mixed with GLC-48 and supplemented with 8:0, 10:0, and
12:0 FAMEs, Nuchek Prep). Concentrations of fatty acids were determined in relation to
peak area of internal standard.
time and tracer on serum lipid class radioactivity were assessed by two-way ANOVA with
unweighted means analysis with comparisons between tracers at each time point done by
independent samples t test. Mean area under the curve was also assessed for each tracer in
each serum lipid class. Comparison of tissue fatty acid concentrations between groups in
PL-DHA versus TAG-DHA supplementation study was done by one-way ANOVA with
Tukey's post-hoc test. Longitudinal changes in DHA concentration in serum lipid classes
were expressed relative to OO-fed rat within each week. Effects of all diets within each
week of supplementation were assessed by one-way ANOVA with Tukey's post-hoc test.
The effects of the different FO, PLC and FO+PLC and time were assessed in longitudinal
serum DHA concentration data by two-way ANOVA.

2.7. PL-DHA vs. TAG-DHA supplementation study calculations 3. Results

The balance method [48,5658] was used to determine brain region, tissue and
whole-body net DHA accretion rates using the following calculation: 3.1. Radiotracer study

Total radioactivity in the brain 6 h after oral gavage of was less than
DHAat beginning of study  DHAat end of study 2 105% of the gavaged dose for both tracers. Rats gavaged with
DHA daily accretion rate 2
number of days of supplementation [U-3H]PC-DHA had a signicantly greater proportion of the radiotracer
94 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102
appear in cerebellum, hippocampus and remainder of brain 6-h
postgavage as compared with rats gavaged with [U-3H]TAG-DHA
(78%, 140% and 69% higher, respectively) with no signicant
differences in cortex, striatum and brainstem (Fig. 1). Unidirectional
incorporation coefcients for tracer uptake from serum were higher in
cerebellum, hippocampus and remainder of brain of [U-3H]PC-DHA
compared with [U-3H]TAG-DHA rats with no differences in other
regions regardless of serum lipid pool examined (data not shown). No
signicant differences in radioactivity were observed in heart, liver,
skin or adipose between [U-3H]TAG-DHA and [U-3H]PC-DHA, al-
though there were differences in the distribution of radioactivity
among liver lipid classes (Supplementary Fig. 1).
Ingestion of both [U- 3H]TAG-DHA and [U-3H]PC-DHA-labeled
serum lipid pools in the rank order TAG N diacylglycerol/cholesterol
N nonesteried fatty acids (NEFA) N PL N monoacylglycerols =
cholesteryl esters (CE). Radioactivity in all fractions appeared to
have plateaued or decreased by the end of the study (Fig. 2). The
highest labeling was observed for serum TAG in the [U-3H]TAG-DHA
gavaged animals at 5-h post gavage with 0.126% of the gavaged dose/
ml of serum. The pattern of labeling of serum lipid pools was different
between [U-3H]TAG-DHA and [U-3H]PC-DHA, as two-way ANOVA
revealed a signicant effect of tracer in the labeling of all serum lipids
pools examined except for total PL (Fig. 2). [U-3H]TAG-DHA tended to
have higher maximum labeling, reach the maximum labeling faster
and maintain a higher value of radioactivity longer compared with
[U-3H]PC-DHA in diacylglycerol/cholesterol, NEFA, monoacylglycerol
and in particular TAG. In contrast, [U-3H]PC-DHA had higher labeling
of CE.
3.2. PL-DHA versus TAG-DHA supplementation study
No differences in body weight or food intake between the dietary
groups were observed over the course of the study (Supplemental Fig.
2). Body weight was 566 53 g at the beginning of supplementation
and 625 64 g at the end of the 4-week supplementation period with
a signicant main effect of time but not of diet. Food intake was 31
5 g/d at the beginning of supplementation and 29 4 g/d at the end of
the study with no changes over time and no differences between diets.
Effects of experimental diets on brain region DHA is shown in Fig. 3
(full brain region fatty acid composition can be found in Supplemental
Tables 27). After 4 weeks of supplementation, rats in the DHA-free
OO supplemented group had brain region total lipid DHA concentra-
Fig. 1. Percent of gavaged radioactivity recovered in total phospholipids of brain regions
6 h after gavage of [U-3H]TAG-DHA and [U-3H]PC-DHA. Signicantly higher enrichment
was found in remainder of brain, cerebellum and hippocampus in rats gavaged with
[U-3H]PC-DHA compared with [U-3H]TAG-DHA. Less than 2105 of the dose was
recovered in total brain phospholipids. Data are mean S.E.M. of 45/group. *:
signicantly different from [U-3H]TAG-DHA gavaged rats by independent samples t test
(Pb.05).
tions of 9.9, 10.1, 10.6, 11.1, 13.4 and 13.7 mol/g in cortex, remainder
of brain, cerebellum, brainstem, hippocampus and striatum, respec-
tively. Supplementation with 2% of dietary fatty acids as DHA in FO,
PLC or FO+PLC resulted in signicantly higher brain DHA concentra-
tions compared with OO-fed controls (3033% higher in cortex, 35
43% higher in cerebellum, 2731% higher in hippocampus, 2832%
higher in brainstem and 2635% higher in remainder of brain) with no
signicant differences in brain DHA concentration between the three
DHA-supplemented groups (Fig. 3). Striatum DHA concentrations
were not signicantly different between groups (P =.053).
The changes in fasting serum lipid class DHA concentrations in
response to DHA supplementation are shown in Fig. 4 (full serum lipid
class fatty acid compositions at conclusion of study are shown in
Supplemental Tables 811). All sources of DHA signicantly increased
all serum lipid DHA concentrations over the course of the study (time
P b .0001 for all lipid classes). The largest increase in DHA concentra-
tion was found in serum triglycerides, with DHA supplementation
raising DHA concentration by 2050-fold, with other fractions
increasing 315-fold. DHA-supplemented rats generally had higher
serum DHA compared to OO controls based on within-week ANOVA
analysis. In two-way ANOVA examining different DHA-containing
diets relative to OO-control, a signicant main effect of DHA
supplementation diet was found only for serum unesteried fatty
acids (P =.0051), in which FO-fed rats had higher DHA at 3 weeks of
supplementation compared with PLC (FO+PLC was intermediate),
while all groups were signicantly higher than control. The only other
instance of signicant differences between DHA-supplemented rats
was in serum TAG after 4 weeks of supplementation, in which
FO+PLC was signicantly higher than FO.
Effects of diets on liver, heart, adipose, intestine and skin DHA
concentration is shown in Fig. 5 (full fatty acid compositions reported
in Supplemental Tables 1216). Supplementation with DHA increased
total lipid DHA concentration by between 7 and 10 fold in liver, heart,
small intestine and skin, while an increase of only three- to fourfold
was found in adipose tissue (P b.05 for all relative to OO). All DHA-
supplemented diets were equally effective at increasing tissue total
lipid DHA concentrations in all tissues except for the heart, in which
the FO+PLC-fed rats had 27% higher heart DHA compared with FO-fed
rats, with PLC being intermediate.
The results of the DHA balance analysis are summarized in Table 2.
Brain region accretion rates were higher in FO, PLC and FO+PLC rats
compared to OO controls and were not different between DHA-
supplemented groups. Whole body DHA was approximately 10-fold
higher in DHA-supplemented groups compared with OO controls with
no differences between DHA sources, resulting in no differences in the
whole body accretion rates. The apparent daily DHA oxidation rate
was highest in FO+PLC-fed rats compared with PLC-fed rats, with FO-
fed being intermediate (all DHA-supplemented groups were higher
than the OO control group), although all DHA-fed groups had similar
proportions of dietary DHA that was apparently -oxidized, ranging
from 84 to 88% of ingested DHA. Among DHA-fed rats, the fecal DHA
excretion was highest in PLC-fed rats compared with FO and FO+PLC
rats, while all DHA-fed rats had higher fecal DHA excretion compared
with OO-fed controls. The percentage of dietary DHA excreted in feces
was higher in PLC-fed rats compared with FO and FO+PLC and was
less than 0.5% of dietary DHA in all groups.
4. Discussion
Numerous groups have developed approaches to target the brain
with DHA [42,5961]; however, whether dietary strategies can also
increase DHA supply to the brain is not clear. Ingestion of PL-DHA has
been suggested to increase blood DHA more effectively compared with
TAG-DHA [3133], and previous studies have shown higher brain
uptake of labeled DHA following ingestion of PL-DHA tracers as
 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102 95

Fig. 2. Percent of gavaged dose recovered in serum lipid classes over 6 h following gavage of [U-3H]TAG-DHA and [U-3H]PC-DHA. Radioactivity of all lipid classes increased during the
experiment, and DHA ester was a signicant factor in all lipid classes except serum phospholipids. [U-3H]TAG-DHA had higher radioactivity in TAG, unesteried fatty acids,
monoacylglycerols, and diacylglycerol/cholesterol while [U-3H]PC-DHA had higher radioactivity in serum cholesteryl esters. Data is mean S.E.M. of three to ve/time-point/group. *:
signicantly different at indicated time point by independent samples t test (Pb.05). AUC: average area under the curve.

compared with TAG-DHA [42,43]. The present study investigated
differences in the acute brain tracer levels and serum lipid pool
enrichment following ingestion of [U-3H]PC-DHA and [U-3H]TAG-
DHA and compared this to the effects of dietary supplementation with
sources of TAG-DHA, PL-DHA or a mixture of both. Similar to previous
studies [42,43], we show that a higher percentage of ingested
[U-3H]PC-DHA radioactivity was found in brain lipids compared
with [U- 3H]TAG-DHA. This effect appears to be specic to the
brain, as no signicant differences in radiotracer enrichment were
found in other tissues examined in general agreement with previous
work [43]. However, the higher brain radiotracer uptake in [U-3H]PC-
DHA gavage did not translate into higher brain DHA concentrations in
PL-DHA fed rats compared with TAG-DHA in the 4-week PL-DHA
versus TAG-DHA supplementation study. This indicates that while
ingestion of DHA esteried to PL may increase delivery of DHA to
brain lipids postprandially, this may not necessarily translate into
96 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102

Fig. 3. DHA concentration of brain regions from rats fed diets supplemented with olive oil, sh oil, sh roe phospholipid concentrate or a mixture of sh oil and sh roe phospholipid
concentrate. All groups supplemented with DHA had higher brain DHA compared with the olive oil control group and were not different from each other. The only exception was
striatum, in which there was no signicant effect of diet on brain DHA concentration. Data are mean S.E.M. for seven to eight/group. Bars with a different superscript are signicantly
different by Tukey's post-hoc test following one-way ANOVA (Pb.05). OO: olive oil, FO: sh oil, PLC: phospholipid concentrate, FO+PLC: sh oil + phospholipid concentrate.

higher brain DHA concentration over the course of a more prolonged
dietary intervention.
There are several possible explanations for the apparent lack of
agreement between the tracer and supplementation experiments. The
DHA supplementation diets used contained 2% of fatty acids as DHA,
which has previously been suggested to be the dietary DHA content at
which brain DHA is maximal [52]. The time-course of DHA repletion
using these diets has not been investigated; however, it is possible that
brain DHA had reached maximal concentration in all DHA-fed groups
by the end of the supplementation period, making it difcult to
observe differences in DHA concentration. It is possible that rats fed
PL-DHA sources (PLC and FO+PLC groups) reached the steady-state
plateau more rapidly compared with FO-fed rats, consistent with
higher uptake of [U-3H]PC-DHA compared with [U-3H]TAG-DHA,
although this potential effect appears to be nullied by extending the
supplementation period to at least 4 weeks. In addition, the striatum
exhibited the lowest DHA accretion rate of any brain region analyzed,
but exhibited no differences in DHA concentration in response to
increasing PL-DHA intake. It is also possible that more DHA is entering
the brain in response to PL-DHA intake, but that it is not resulting in
increased accretion of DHA. For example, a previous study from our
group showed that rats fed ALA- or DHA-containing diets have similar
brain DHA concentrations despite DHA-fed animals having higher
brain DHA uptake rates [48]. In these cases, the turnover of brain DHA
may be increased; however, this requires investigation using in vivo
kinetic modeling [50]. A greater uptake of a tracer into brain lipids also
may not represent a major quantitative contribution to DHA
homeostasis and, as a result, may not translate into higher brain
DHA concentration. However, it is possible that the higher brain tracer
concentration in [U-3H]PC-DHA compared with [U-3H]TAG-DHA rats
indicates more rapid DHA uptake, which may be advantageous for
rapid delivery of DHA in certain clinical settings [62]. Obese Zucker rats
fed krill oil, a PL-DHA source, have higher brain PL DHA compared to
sh oil-fed rats [63], possibly suggesting a selective effect of PL-DHA
on brain DHA in obese animals that may not occur in healthy animals.
The mechanism underlying the higher brain radioactivity in
[U-3H]PC-DHA compared with [U-3H]TAG-DHA rats is unclear, but
may involve differences in kinetics of serum radiotracer enrichment.
Slightly higher labeling of serum unesteried fatty acids, the serum
lipid pool that appears to be the main pool for brain DHA uptake
 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102 97

Fig. 4. DHA concentration of serum lipid fractions from rats fed diets supplemented with OO, FO, PLC or FO+PLC expressed relative to olive oil-supplemented rats. DHA supplementation
increased serum lipid class DHA concentration in all DHA-supplemented groups relative to baseline. The source of DHA was only a signicant factor in the DHA concentration of serum
unesteried fatty acids, in which PLC-supplemented rats had lower DHA than FO rats after 21 days of supplementation, but this difference disappeared at 28 days. FO+PLC had higher
serum TAG DHA at 28 days compared with FO. Data are mean S.E.M. for seven to eight/group. Bars with a different superscript within a time point are signicantly different within that
time point by Tukey's post-hoc test following one-way ANOVA (Pb.05). Indicated P-values are derived from two-way ANOVA of time and DHA ester. OO: olive oil, FO: sh oil, PLC:
phospholipid concentrate, FO+PLC: sh oil + phospholipid concentrate.

[25,27], was observed in rats gavaged with [U-3H]TAG-DHA, which is
not in agreement with the lower brain tracer uptake in [U-3H]TAG-
DHA rats relative to [U-3H]PC-DHA. [U-3H]PC-DHA more effectively
labeled serum cholesteryl esters, consistent with preferential labeling
of HDL that has been observed for dietary PL previously [44,45]. It has
been proposed that small HDL particles can enter the brain through
the action of scavenger receptor Class B, Type I (SR-BI) on brain
capillary endothelial cells [6466], suggesting a possible mechanism
for the higher brain tracer uptake in [U-3H]PC-DHA-gavaged rats
compared with [U-3H]TAG-DHA. SR-BI mediates cholesteryl ester
inux in the liver [67], and as such may be involved in brain cholesteryl
ester uptake in the brain as well. Brain HDL uptake has previously been
proposed to be involved in brain DHA uptake [68], and HDL-associated
radioactivity is detected in mouse brain following injection of labeled
HDL particles [69]. Maintenance of whole-brain DHA concentration
does not require the low-density or very low density lipoprotein
receptors [70,71]; however, the role of HDL in brain DHA homeostasis
and targeted delivery of DHA to the brain is not clear.
The postprandial distribution of [U-3H]TAG-DHA and [U-3H]PC-
DHA radioactivity among the different serum lipid pools likely reect
the differences in their digestion and absorption. Dietary TAG are
predominantly hydrolyzed by pancreatic lipase, which primarily
hydrolyzes fatty acids in the sn-1 and 3 positions of TAG, resulting
in two fatty acids and a sn-2-monoacylglycerol (MAG) [72]. These
98 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102

Fig. 5. DHA concentration of tissues from rats fed diets supplemented with OO, FO, PLC and FO+PLC. All DHA-containing diets had higher tissue DHA compared with OO controls.
FO+PLC rats had higher heart DHA compared with FO-fed rats, and DHA concentrations in the rest of the tissues examined were not different among DHA-supplementation groups. Data
are mean S.E.M. for seven to eight/group. Bars with a different superscript are signicantly different by Tukey's post-hoc test following one-way ANOVA (Pb.0.05). OO: olive oil, FO:
sh oil, PLC: phospholipid concentrate, FO+PLC: sh oil + phospholipid concentrate.

products are taken up into the enterocyte, and the 2-MAG is utilized
for TAG synthesis for chylomicron secretion [73]. Therefore, at least
one third of the acyl-chain radioactivity of [U-3H]TAG-DHA is utilized
for TAG synthesis, and this is reected in the radioactive enrichment of
serum TAG and its metabolites, MAG and DAG following [U-3H]TAG-
DHA ingestion. As the DHA in the [U-3H]TAG-DHA is in the sn-2
position, it is likely that the majority is absorbed as 2-DHA-MAG and
reesteried in the enterocyte. PL, on the other hand, is hydrolyzed by
activities of phospholipase A1 (pancreatic phospholipase A1[74],
secondary activity of pancreatic lipase in some species [75]) and A2
(pancreatic phospholipase A2[76], pancreatic lipase-related protein 2
[77]) into free fatty acids and sn-1/2-lysoPL. In the rat, approximately
half of the acyl chain moiety of ingested PC is absorbed as lyso-PC and
the other half as unesteried fatty acids, with the dominant LPC
species being 1-acyl-lysoPC [78], indicating that the major intestinal
phospholipase activity in rats is phospholipase A2. As the DHA in the
[U-3H]PC-DHA tracer was in the sn-2 position, it is likely that it was
absorbed in the unesteried form. [U-3H]PC-DHA appears to be a
poorer substrate for enterocytic TAG synthesis as compared with
[U-3H]TAG-DHA, and the more effective labeling of serum cholesteryl
esters may result from the preferential incorporation of acyl moieties
from ingested PL into HDL observed previously [44,45]. [U-3H]PC-
DHA-gavaged rats also had a higher proportion of liver radioactivity in
the cholesteryl ester fraction, suggesting that secretion of radioactive
CE by the liver may have been higher in these rats compared to
[U-3H]TAG-DHA, although serum lecithin-cholesterol acyltransferase
may also be involved. It has been suggested that sn-1-lyso-2-DHA-PC
is the main serum DHA pool supplying the brain [30,61]; however, we
did not observe any differences in serum phospholipid labeling
between the two tracers, and fasting serum lysoPC DHA concentration
was similar between the three DHA-supplemented groups. The
presence of PL in PL-DHA may increase intestinal absorption of lipids
as biliary PC is required for intestinal TAG absorption [79] and
inclusion of PC in intestinal lipid infusion increases TAG absorption
[7981]. However, similar whole-body DHA accretion rates between
FO, PLC and FO+PLC rats indicate that inclusion of PL-DHA in DHA
 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102
Table 2
Summary of DHA balance analysis
Olive oil Fish oil
Whole-body DHA after 4 weeks (mol) 685a 68057b
Total dietary DHA intake (mol) 14620a 54021147b
Whole body DHA accretion (mol) 235a 63657b
Whole body DHA accretion rate (mol/d) 0.80.2a 22.72.0b
Metabolic DHA consumption rate (mol/d) 31a 16816bc
% Apparent Dietary DHA -oxidized 524a 871b
Total fecal DHA excretion (mol) 0.60.1a 9.62.1b
% Dietary DHA excreted in feces 0.430.10ab 0.180.04a
Cortex DHA accretion rate (nmol/g/d) 1518a 1315b
Cerebellum DHA accretion rate (nmol/g/d) 319a 10124b
Hippocampus DHA accretion rate (nmol/g/d) 925a 14031b
Striatum DHA accretion rate (nmol/g/d) 31105 346
Brainstem DHA accretion rate (nmol/g/d) 1134a 10110b
Remainder of brain DHA accretion rate (nmol/g/d) 2910a 7312b
Data are mean S.E.M. n=78/group. After 11-week n-3 PUFA deprivation, 8 rats were sacriced to measure baseline DHA concentrations, and the difference after 4 weeks of
supplementation with DHA or control diet were determined to measure accretion rates. At baseline, whole-body DHA was 453 mol (n=8). Different superscript letters that indicate
means are signicantly different (Pb.05) determined by Tukey's post-hoc comparison following one-way ANOVA.
formulations does not affect the amount of DHA absorbed, and in fact
the PLC-fed rats had the highest fecal DHA concentration (albeit less
than 0.5% of dietary DHA). However, a more direct measure of
intestinal DHA absorption is needed to denitively conclude this.
Importantly, the differences in postprandial serum lipid radioac-
tivity following gavage of [U-3H]TAG-DHA and [U-3H]PC-DHA did
not translate into clear differences in fasting serum lipid DHA
concentration in response to supplementing with a PL- compared
with TAG-DHA source for 4 weeks. Differences in serum lipid DHA
among the PL- and TAG-DHA-supplemented groups were time-point
specic, and PL- versus TAG-DHA source was not a signicant factor for
any lipid class except for unesteried fatty acids. This suggests that
analysis of postprandial absorption kinetics will not match fasting
blood sampling following an intervention period. A recent randomized
controlled trial comparing krill oil (PL-EPA + DHA source) and sh oil
(TAG-EPA+DHA source) shows that krill oil had higher postprandial
serum phospholipid EPA+DHA, but that this trend was reversed
(albeit not statistically signicant) when examining plasma TAG
EPA+DHA [82]. Thus, the lipid pool examined can have signicant
effects on the conclusions of DHA bioavailability studies, indicating
that such studies should, in the absence of any clear rationale for
investigating a particular serum lipid class, examine bioavailability of
multiple serum lipid pools or of serum total lipids. Some of these
factors have been previously proposed to introduce signicant
heterogeneity in the assessment of DHA bioavailability [39].
The mixture of the FO and PLC DHA sources appeared to have a
positive effect on some measures of DHA concentration. The FO+PLC
group (in which approximately 30% of the total DHA was esteried to
phospholipids) had higher heart DHA and serum TAG DHA compared
with the FO group, consistent with higher heart DHA in krill oil- versus
sh oil-fed obese Zucker rats [83]. Interestingly, the higher heart DHA
in the FO+PLC rats relative to the FO rats did not correspond to
differences in heart radioactivity between [U- 3H]PC-DHA and
[U-3H]TAG-DHA rats, suggesting that this effect is due to either the
specic mixture of PL- and TAG-DHA or the 4 weeks of supplemen-
tation. In the heart, DHA is also extensively incorporated into
cardiolipin [84] which is associated with modications to mitochon-
drial enzyme function [85]. The acute absorption and tissue uptake
kinetics of various mixtures of various ratios PL-DHA and TAG-DHA
warrant investigation, as do the potential cardioprotective effects
specic to certain mixtures of PL-DHA and TAG-DHA.
The amount of tracer that appeared in the brain for both [U-3H]PC-
DHA and [U-3H]TAG-DHA tracers was low (less than 2 105% of the
dose) but is roughly consistent with previous studies (approximately
0.2% and 0.4% of dose in after 24 h in rat whole brain and 6 days in
99
Phospholipid concentrate Fish oil + phospholipid concentrate
79762b 68964b
5141803b 6145454b
75262b 64426b
26.82.2b 23.02.3b
1589b 1946c
841b 881b
23.01.6c 11.60.9b
0.460.04b 0.190.1a
1284b 1249b
9511b 1126b
12113b 13726b
369 5417
10311b 11515b
6713b 964b
piglet gray matter, respectively[42, 43]) considering differences in
sampling time postgavage, species, age of animals, analytical meth-
odologies and diet between the studies. This suggests that only a
minor fraction of ingested DHA targets the brain, regardless of the
form in which it is consumed. The major fate of PL-DHA and TAG-DHA
is oxidation, as evidenced by the high proportion of the DHA
consumed being -oxidized (8488%), consistent with previous
studies showing high DHA oxidation using the balance method to
examine DHA balance when fed as microalgal oil [58] or ethyl ester
[48], and the high proportion of label recovered in breath following
ingestion of [U-13C]DHA following sh oil supplementation [86].
Interestingly, the OO-fed rats had the lowest proportion of dietary
DHA oxidized (52%), indicating that DHA may be selectively retained
in these animals, consistent with previous work demonstrating
extended DHA half-life in brain following 15-week n-3 PUFA
deprivation [46] and lower breath recovery of [U-13C]DHA label in
individuals pre- compared with postsh oil supplementation. A higher
rate of apparent metabolic DHA consumption (i.e., apparent -
oxidation) was observed in rats fed the FO+PLC diet compared to
PLC, but this appears to have resulted from slightly higher intake of
DHA, as the percent of ingested DHA that was -oxidized was not
different between groups. Krill-oil fed rats have recently been shown
to accrete a higher proportion of dietary DHA in whole body compared
with sh oil-fed rats after 6 weeks of supplementation [87]. No
signicant differences in whole-body DHA accretion were observed in
the present study of 4-week supplementation; however, PL-fed rats
had nonsignicantly higher whole-body DHA accretion compared to
FO and FO+PLC-fed rats. It is possible that longer duration of
supplementation is required to elicit differences in whole-body DHA
accretion between PL- and TAG-DHA.
There are several considerations regarding the interpretation of
this work. The isotope used was uniformly labeled, and as such, the
conclusions rendered from this study pertain to the entire sn-1,3-
palmitoyl-2-docosahexaenoylglycerol or 1-palmitoyl-2-docosahex-
aenoyl-sn-glycero-3-phosphocholine and not necessarily the DHA
contained within them. In [U-3H]TAG-DHA, the acyl chain radioactiv-
ity is one-third DHA and two-thirds palmitate, while in [U-3H]PC-DHA
it is one-half DHA and one-half palmitate. This likely does not affect
the serum lipid labeling of the tracers, as the appearance of DHA and
palmitate in serum lipids is very similar when provided orally as
unesteried fatty acids (TAGNNNEFANPLNDAGNCE for DHA [88] and
TAGNNNEFANCENPLNDAG for palmitate [27]). In addition, little
conversion of DHA to other fatty acids is likely to have occurred in
this study, as other studies have found low enrichment of other fatty
acids with label provided as DHA 22 h after oral administration in liver
100 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102
in suckling rats [89] or 5 h after intravenous administration in plasma
in maternal baboons [90]. Similarly, appearance of orally administered
palmitate label as 18:0 and 16:1 is approximately 8% after 7 days in
humans [91] and has been below the detection limit 24 h after oral
dose in humans in one study [92], suggesting that the majority of
labeled palmitate ingested remains as labeled palmitate over the
course of this study. The higher brain radioactivity resulting from
[U-3H]PC-DHA compared with [U-3H]TAG-DHA gavage presently is
consistent with previous work in which only the sn-2 DHA is labeled
[42,43], indicating that this higher brain radioactivity in [U-3H]PC-
DHA rats was, at least in part, due to higher brain uptake of the
[3H]DHA moiety. Interestingly, baboon neonate brain takes up more
arachidonate when fed as sn-2-arachidonate-PC compared with sn-2-
arachidonate-TAG [93], suggesting that this effect is not limited to sn-
2-DHA-PC. The tracers were both dissolved in olive oil (virtually
entirely TAG) prior to gavage, and as such, the absorption matrix for
both TAG and PL tracers was TAG. This may have had some effects on
the absorption kinetics of the PL tracer. The DAG and cholesterol
fractions were collected together in the radiotracer gavage study, and
so the distribution of radioactivity between these two fractions is
unclear as carbon recycling into cholesterol could account for some of
the radioactivity. However, the appearance of carbon in serum saturated
fatty acids is low for DHA in rats over this time course (b 3% of serum
label) [90], suggesting that cholesterol labeling via carbon recycling of
labeled DHA was likely low in this study. Another factor is the baseline
diet used in these studies, which is a low n-3 PUFA diet that has
previously been shown to increase brain DHA half-life [46] and hepatic
synthesis rate [94]. The present studies therefore represent n-3 PUFA
repletion on a very low n-3 PUFA background. However, the association
between low n-3 PUFA status and neurodegenerative [711] and
neuropsychiatric disorders [1316] and cardiovascular mortality [95]
suggests that this experimental approach is relevant for determining
the relative efcacy of different dietary forms of DHA on tissue DHA
uptake and concentration. Also, as the diet is essentially free of ALA
(b 0.2% of dietary fatty acids), DHA synthesis from dietary ALA likely
does not contribute appreciably to changes in tissue and blood DHA
status throughout the study between DHA-supplemented groups [48].
The three DHA-supplemented groups were matched for DHA intake;
however, there were small differences in the EPA and 22:5n-3 dietary
content. These differences amounted to a range in total dietary n-3 PUFA
content between the three DHA-supplemented groups of only 3.31
3.41% of total fatty acids.
This study conrms greater brain uptake following an oral tracer
dose of DHA esteried to PL compared to TAG [42,43] but shows that
supplementing diets with a source of PL-DHA, TAG-DHA or a mixture
of both results in similar increases in brain DHA compared to a low n-3
diet. This suggests that the higher tracer uptake of PL-DHA compared
to TAG-DHA does not necessarily translate into higher brain DHA
concentration. In addition, signicant differences were observed in the
postprandial serum lipid radioactivity prole following gavage of
[U-3H]PC-DHA compared with [U-3H]TAG-DHA, while only time
point-specic differences were observed in fasting serum DHA
concentrations during the 4-week PL-DHA versus TAG-DHA supple-
mentation study. This indicates that the study and sampling protocol
can have signicant effects on the conclusions obtained from studies
evaluating the effects of PL-DHA compared with TAG-DHA on blood
DHA levels. Overall, there appear to be differences in the digestion,
absorption and brain uptake of DHA esteried to PL compared to TAG;
however, these postprandial differences may not translate into
increased longer term blood and tissue DHA status following
supplementation with PL-DHA compared with TAG-DHA.
Potential conflict of interest
This study was funded by Arctic Nutrition AS, NO-6155 rsta, Norway.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jnutbio.2016.02.009.
References
[1] Igarashi M, Ma K, Gao F, Kim HW, Greenstein D, Rapoport SI, et al. Brain lipid
concentrations in bipolar disorder. J Psychiatr Res 2010;44:17782.
[2] Rao JS, Ertley RN, Lee HJ, DeMar Jr JC, Arnold JT, Rapoport SI, et al. n-3
polyunsaturated fatty acid deprivation in rats decreases frontal cortex BDNF via a
p38 MAPK-dependent mechanism. Mol Psychiatry 2007;12:3646.
[3] Cao D, Kevala K, Kim J, Moon HS, Jun SB, Lovinger D, et al. Docosahexaenoic acid
promotes hippocampal neuronal development and synaptic function. J Neuro-
chem 2009;111:51021.
[4] Su HM. Mechanisms of n-3 fatty acid-mediated development and maintenance of
learning memory performance. J Nutr Biochem 2010;21:36473.
[5] Orr SK, Palumbo S, Bosetti F, Mount HT, Kang JX, Greenwood CE, et al. Unesterified
docosahexaenoic acid is protective in neuroinflammation. J Neurochem 2013;
127:37893.
[6] Brenna JT. Animal studies of the functional consequences of suboptimal
polyunsaturated fatty acid status during pregnancy, lactation and early post-
natal life. Matern Child Nutr 2011;7(Suppl. 2):5979.
[7] Igarashi M, Ma K, Gao F, Kim HW, Rapoport SI, Rao JS. Disturbed choline
plasmalogen and phospholipid fatty acid concentrations in Alzheimer's disease
prefrontal cortex. J Alzheimers Dis 2011;24:50717.
[8] Soderberg M, Edlund C, Kristensson K, Dallner G. Fatty acid composition of brain
phospholipids in aging and in Alzheimer's disease. Lipids 1991;26:4215.
[9] Lukiw WJ, Cui JG, Marcheselli VL, Bodker M, Botkjaer A, Gotlinger K, et al. A role for
docosahexaenoic acid-derived neuroprotectin D1 in neural cell survival and
Alzheimer disease. J Clin Invest 2005;115:277483.
[10] Astarita G, Jung KM, Berchtold NC, Nguyen VQ, Gillen DL, Head E, et al. Deficient
liver biosynthesis of docosahexaenoic acid correlates with cognitive impairment
in Alzheimer's disease. PLoS One 2010;5, e12538.
[11] Cunnane SC, Schneider JA, Tangney C, Tremblay-Mercier J, Fortier M, Bennett DA,
et al. Plasma and brain fatty acid profiles in mild cognitive impairment and
Alzheimer's disease. J Alzheimers Dis 2012;29:6917.
[12] Cunnane SC, Chouinard-Watkins R, Castellano CA, Barberger-Gateau P. Docosa-
hexaenoic acid homeostasis, brain aging and Alzheimer's disease: can we
reconcile the evidence? Prostaglandins Leukot Essent Fatty Acids 2013;88:6170.
[13] McNamara RK, Jandacek R, Tso P, Dwivedi Y, Ren X, Pandey GN. Lower
docosahexaenoic acid concentrations in the postmortem prefrontal cortex of
adult depressed suicide victims compared with controls without cardiovascular
disease. J Psychiatr Res 2013;47:118791.
[14] McNamara RK, Hahn CG, Jandacek R, Rider T, Tso P, Stanford KE, et al. Selective
deficits in the omega-3 fatty acid docosahexaenoic acid in the postmortem
orbitofrontal cortex of patients with major depressive disorder. Biol Psychiatry
2007;62:1724.
[15] McNamara RK, Jandacek R, Rider T, Tso P, Stanford KE, Hahn CG, et al. Deficits in
docosahexaenoic acid and associated elevations in the metabolism of arachidonic
acid and saturated fatty acids in the postmortem orbitofrontal cortex of patients
with bipolar disorder. Psychiatry Res 2008;160:28599.
[16] McNamara RK, Jandacek R, Rider T, Tso P, Hahn CG, Richtand NM, et al.
Abnormalities in the fatty acid composition of the postmortem orbitofrontal
cortex of schizophrenic patients: gender differences and partial normalization
with antipsychotic medications. Schizophr Res 2007;91:3750.
[17] Hamazaki K, Hamazaki T, Inadera H. Abnormalities in the fatty acid composition of
the postmortem entorhinal cortex of patients with schizophrenia, bipolar
disorder, and major depressive disorder. Psychiatry Res 2013;210:34650.
[18] Hamazaki K, Hamazaki T, Inadera H. Fatty acid composition in the postmortem
amygdala of patients with schizophrenia, bipolar disorder, and major depressive
disorder. J Psychiatr Res 2012;46:10248.
[19] Fraser T, Tayler H, Love S. Fatty acid composition of frontal, temporal and parietal
neocortex in the normal human brain and in Alzheimer's disease. Neurochem Res
2010;35:50313.
[20] Oster T, Pillot T. Docosahexaenoic acid and synaptic protection in Alzheimer's
disease mice. Biochim Biophys Acta 2010;1801:7918.
[21] Cole GM, Frautschy SA. Docosahexaenoic acid protects from amyloid and
dendritic pathology in an Alzheimer's disease mouse model. Nutr Health 2006;
18:24959.
[22] Zugno AI, Chipindo HL, Volpato AM, Budni J, Steckert AV, de Oliveira MB, et al.
Omega-3 prevents behavior response and brain oxidative damage in the ketamine
model of schizophrenia. Neuroscience 2014;259:22331.
[23] Park Y, Moon HJ, Kim SH. N-3 polyunsaturated fatty acid consumption produces
neurobiological effects associated with prevention of depression in rats after the
forced swimming test. J Nutr Biochem 2012;23:9248.
[24] Igarashi M, DeMar Jr JC, Ma K, Chang L, Bell JM, Rapoport SI. Docosahexaenoic acid
synthesis from alpha-linolenic acid by rat brain is unaffected by dietary n-3 PUFA
deprivation. J Lipid Res 2007;48:11508.
[25] Moore SA, Yoder E, Murphy S, Dutton GR, Spector AA. Astrocytes, not neurons,
produce docosahexaenoic acid (22:6 omega-3) and arachidonic acid (20:4
omega-6). J Neurochem 1991;56:51824.
 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102
[26] Contreras MA, Greiner RS, Chang MC, Myers CS, Salem Jr N, Rapoport SI.
Nutritional deprivation of alpha-linolenic acid decreases but does not abolish
turnover and availability of unacylated docosahexaenoic acid and docosahex-
aenoyl-CoA in rat brain. J Neurochem 2000;75:2392400.
[27] Purdon D, Arai T, Rapoport S. No evidence for direct incorporation of esterified
palmitic acid from plasma into brain lipids of awake adult rat. J Lipid Res 1997;38:
52630.
[28] Chen CT, Kitson AP, Hopperton KE, Domenichiello AF, Lin L, Trepanier MO, et al.
Plasma non-esterified docosahexaenoic acid is the major pool supplying the brain.
Sci Rep 2015 [in press].
[29] Nguyen LN, Ma D, Shui G, Wong P, Cazenave-Gassiot A, Zhang X, et al. Mfsd2a is a
transporter for the essential omega-3 fatty acid docosahexaenoic acid. Nature
2014;509.
[30] Thies F, Pillon C, Moliere P, Lagarde M, Lecerf J. Preferential incorporation of sn-2
lysoPC DHA over unesterified DHA in the young rat brain. Am J Physiol 1994;267:
R12739.
[31] Ramprasath VR, Eyal I, Zchut S, Jones PJ. Enhanced increase of omega-3 index in
healthy individuals with response to 4-week n-3 fatty acid supplementation from
krill oil versus fish oil. Lipids Health Dis 2013;12:178.
[32] Ulven SM, Kirkhus B, Lamglait A, Basu S, Elind E, Haider T, et al. Metabolic effects of
krill oil are essentially similar to those of fish oil but at lower dose of EPA and DHA,
in healthy volunteers. Lipids 2011;46:3746.
[33] Maki KC, Reeves MS, Farmer M, Griinari M, Berge K, Vik H, et al. Krill oil
supplementation increases plasma concentrations of eicosapentaenoic and
docosahexaenoic acids in overweight and obese men and women. Nutr Res
2009;29:60915.
[34] Schuchardt JP, Schneider I, Meyer H, Neubronner J, von Schacky C, Hahn A.
Incorporation of EPA and DHA into plasma phospholipids in response to different
omega-3 fatty acid formulationsa comparative bioavailability study of fish oil vs.
krill oil. Lipids Health Dis 2011;10:145.
[35] Vaisman N, Kaysar N, Zaruk-Adasha Y, Pelled D, Brichon G, Zwingelstein G, et al.
Correlation between changes in blood fatty acid composition and visual sustained
attention performance in children with inattention: effect of dietary n-3 fatty
acids containing phospholipids. Am J Clin Nutr 2008;87:117080.
[36] Carnielli VP, Verlato G, Pederzini F, Luijendijk I, Boerlage A, Pedrotti D, et al.
Intestinal absorption of long-chain polyunsaturated fatty acids in preterm infants
fed breast milk or formula. Am J Clin Nutr 1998;67:97103.
[37] Yurko-Mauro K, Kralovec J, Bailey-Hall E, Smeberg V, Stark JG, Salem Jr N. Similar
eicosapentaenoic acid and docosahexaenoic acid plasma levels achieved with fish
oil or krill oil in a randomized double-blind four-week bioavailability study. Lipids
Health Dis 2015;14:99.
[38] Salem Jr N, Kuratko CN. A reexamination of krill oil bioavailability studies. Lipids
Health Dis 2014;13:137.
[39] Ghasemifard S, Turchini GM, Sinclair AJ. Omega-3 long chain fatty acid
"bioavailability": a review of evidence and methodological considerations. Prog
Lipid Res 2014;56:92108.
[40] Nichols PD, Kitessa SM, Abeywardena M. Commentary on a trial comparing krill
oil versus fish oil. Lipids Health Dis 2014;13:2.
[41] Ramprasath VR, Eyal I, Zchut S, Jones PJ. Response to commentary on a trial
comparing krill oil versus fish oil. Lipids Health Dis 2014;13:17.
[42] Liu L, Bartke N, Van Daele H, Lawrence P, Qin X, Park HG, et al. Higher efficacy of
dietary DHA provided as a phospholipid than as a triglyceride for brain DHA
accretion in neonatal piglets. J Lipid Res 2014;55:5319.
[43] Graf BA, Duchateau GS, Patterson AB, Mitchell ES, van Bruggen P, Koek JH, et al.
Age dependent incorporation of 14C-DHA into rat brain and body tissues after
dosing various 14C-DHA-esters. Prostaglandins Leukot Essent Fatty Acids 2010;
83:8996.
[44] Zierenberg O, Grundy SM. Intestinal absorption of polyenephosphatidylcholine in
man. J Lipid Res 1982;23:113642.
[45] Amate L, Gil A, Ramirez M. Feeding infant piglets formula with long-chain
polyunsaturated fatty acids as triacylglycerols or phospholipids influences the
distribution of these fatty acids in plasma lipoprotein fractions. J Nutr 2001;131:
12505.
[46] DeMar Jr JC, Ma K, Bell JM, Rapoport SI. Half-lives of docosahexaenoic acid in rat
brain phospholipids are prolonged by 15 weeks of nutritional deprivation of n-3
polyunsaturated fatty acids. J Neurochem 2004;91:112537.
[47] Folch J, Lees M, Sloane Stanley GHS. A simple method for the isolation and
purification of total lipides from animal tissues. J Biol Chem 1957;226:497509.
[48] Domenichiello AF, Chen CT, Trepanier MO, Stavro PM, Bazinet RP. Whole body
synthesis rates of docosahexaenoic acid (DHA) from alpha-linolenic acid are
greater than brain DHA accretion and uptake rates in adult rats. J Lipid Res 2013;
55:6274.
[49] Chen CT, Domenichiello AF, Trepanier MO, Liu Z, Masoodi M, Bazinet RP. The low
levels of eicosapentaenoic acid in rat brain phospholipids are maintained via
multiple redundant mechanisms. J Lipid Res 2013;54:241022.
[50] Robinson PJ, Noronha J, DeGeorge JJ, Freed LM, Nariai T, Rapoport SI. A
quantitative method for measuring regional in vivo fatty-acid incorporation
into and turnover within brain phospholipids: review and critical analysis. Brain
Res Brain Res Rev 1992;17:187214.
[51] Lin LE, Chen CT, Hildebrand KD, Liu Z, Hopperton KE, Bazinet RP. Chronic dietary
n-6 PUFA deprivation leads to conservation of arachidonic acid and more rapid
loss of DHA in rat brain phospholipids. J Lipid Res 2015;56:390402.
[52] Orr SK, Tong JY, Kang JX, Ma DW, Bazinet RP. The fat-1 mouse has brain
docosahexaenoic acid levels achievable through fish oil feeding. Neurochem Res
2010;35:8119.
101
[53] Wang WQ, Gustafson A. One-dimensional thin-layer chromatographic separation
of phospholipids and lysophospholipids from tissue lipid extracts. J Chromatogr
1992;581:13942.
[54] Morrison WR, Smith LM. Preparation of fatty acid methyl esters and dimethy-
lacetals from lipids with boron fluoride-methanol. J Lipid Res 1964;4:6008.
[55] Chen CT, Liu Z, Bazinet RP. Rapid de-esterification and loss of eicosapentaenoic
acid from rat brain phospholipids: an intracerebroventricular study. J Neurochem
2011;116:36373.
[56] Yang J, Chen ZY, Cunnane SC. Application of the balance method to determining
accumulation, metabolism, and apparent oxidation of linoleic and alpha-linolenic
acids in the pregnant rat. Metabolism 1994;43:9404.
[57] Cunnane SC, Anderson MJ. The majority of dietary linoleate in growing rats is
beta-oxidized or stored in visceral fat. J Nutr 1997;127:14652.
[58] Poumes-Ballihaut C, Langelier B, Houlier F, Alessandri JM, Durand G, Latge C, et al.
Comparative bioavailability of dietary alpha-linolenic and docosahexaenoic acids
in the growing rat. Lipids 2001;36:793800.
[59] Belayev L, Khoutorova L, Atkins KD, Eady TN, Hong S, Lu Y, et al. Docosahexaenoic
acid therapy of experimental ischemic stroke. Transl Stroke Res 2011;2:3341.
[60] Hachem M, Geloen A, Van AL, Foumaux B, Fenart L, Gosselet F, et al. Efficient
Docosahexaenoic Acid Uptake by the Brain from a Structured Phospholipid. Mol
Neurobiol 2015. http://dx.doi.org/10.1007/s12035-015-9228-9.
[61] Lagarde M, Bernoud N, Brossard N, Lemaitre-Delaunay D, Thies F, Croset M, et al.
Lysophosphatidylcholine as a preferred carrier form of docosahexaenoic acid to
the brain. J Mol Neurosci 2001;16:2014 [discussion 15-21].
[62] Barrett EC, McBurney MI, Ciappio ED. Omega-3 fatty acid supplementation as a
potential therapeutic aid for the recovery from mild traumatic brain injury/
concussion. Adv Nutr 2014;5:26877.
[63] Di Marzo V, Grinari M, Carta G, Murru E, Ligresti A, Cordeddu L, et al. Dietary krill
oil increases docosahexaenoic acid and reduces 2-arachidonoylglycerol but not N-
acylethanolamine levels in the brain of obese Zucker rats. Int Dairy J 2010;20:
2315.
[64] Goti D, Hrzenjak A, Levak-Frank S, Frank S, van der Westhuyzen DR, Malle E, et al.
Scavenger receptor class B, type I is expressed in porcine brain capillary
endothelial cells and contributes to selective uptake of HDL-associated vitamin
E. J Neurochem 2001;76:498508.
[65] Vitali C, Wellington CL, Calabresi L. HDL and cholesterol handling in the brain.
Cardiovasc Res 2014;103:40513.
[66] Panzenboeck U, Balazs Z, Sovic A, Hrzenjak A, Levak-Frank S, Wintersperger A,
et al. ABCA1 and scavenger receptor class B, type I, are modulators of reverse
sterol transport at an in vitro bloodbrain barrier constituted of porcine brain
capillary endothelial cells. J Biol Chem 2002;277:427819.
[67] Acton S, Rigotti A, Landschulz KT, Xu S, Hobbs HH, Krieger M. Identification of
scavenger receptor SR-BI as a high density lipoprotein receptor. Science 1996;
271:51820.
[68] Polozova A, Salem Jr N. Role of liver and plasma lipoproteins in selective transport
of n-3 fatty acids to tissues: a comparative study of 14C-DHA and 3H-oleic acid
tracers. J Mol Neurosci 2007;33:5666.
[69] Polozova A, Gionfriddo E, Salem Jr N. Effect of docosahexaenoic acid on tissue
targeting and metabolism of plasma lipoproteins. Prostaglandins Leukot Essent
Fatty Acids 2006;75:18390.
[70] Rahman T, Taha AY, Song BJ, Orr SK, Liu Z, Chen CT, et al. The very low
density lipoprotein receptor is not necessary for maintaining brain polyunsatu-
rated fatty acid concentrations. Prostaglandins Leukot Essent Fatty Acids 2010;82:
1415.
[71] Chen CT, Ma DW, Kim JH, Mount HT, Bazinet RP. The low density lipoprotein
receptor is not necessary for maintaining mouse brain polyunsaturated fatty acid
concentrations. J Lipid Res 2008;49:14752.
[72] Tso P, Hayashi H. The physiology and regulation of the intestinal absorption and
transport of omega-3 and omega-6 fatty acids. Adv Prostaglandin Thromboxane
Leukot Res 1989;19:6236.
[73] Schuchardt JP, Hahn A. Bioavailability of long-chain omega-3 fatty acids.
Prostaglandins Leukot Essent Fatty Acids 2013;89:18.
[74] Arima N, Inoue A, Makide K, Nonaka T, Aoki J. Surface loops of extracellular
phospholipase A(1) determine both substrate specificity and preference for
lysophospholipids. J Lipid Res 2012;53:51321.
[75] Aloulou A, Frikha F, Noiriel A, Bou Ali M, Abousalham A. Kinetic and structural
characterization of triacylglycerol lipases possessing phospholipase A1 activity.
Biochim Biophys Acta 1841;2014:5817.
[76] Ono T, Tojo H, Inoue K, Kagamiyama H, Yamano T, Okamoto M. Rat pancreatic
phospholipase A2: purification, characterization, and N-terminal amino acid
sequence. J Biochem (Tokyo) 1984;96:78592.
[77] Roussel A, Yang Y, Ferrato F, Verger R, Cambillau C, Lowe M. Structure and activity
of rat pancreatic lipase-related. protein 2. J Biol Chem 1998;273:321218.
[78] Le Kim D, Betzing H. Intestinal absorption of polyunsaturated phosphatidylcho-
line in the rat. Hoppe Seylers Z Physiol Chem 1976;357:132131.
[79] Tso P, Kendrick H, Balint JA, Simmonds WJ. Role of biliary phosphatidylcholine in
the absorption and transport of dietary triolein in the rat. Gastroenterology 1981;
80:605.
[80] Mansbach II CM, Arnold A, Cox MA. Factors influencing triacylglycerol delivery
into mesenteric lymph. Am J Physiol 1985;249:G6428.
[81] Nishimukai M, Wakisaka T, Hara H. Ingestion of plasmalogen markedly increased
plasmalogen levels of blood plasma in rats. Lipids 2003;38:122735.
[82] Kohler A, Sarkkinen E, Tapola N, Niskanen T, Bruheim I. Bioavailability of fatty
acids from krill oil, krill meal and fish oil in healthy subjectsa randomized, single-
dose, cross-over trial. Lipids Health Dis 2015;14:19.
102 A.P. Kitson et al. / Journal of Nutritional Biochemistry 33 (2016) 91102
[83] Batetta B, Griinari M, Carta G, Murru E, Ligresti A, Cordeddu L, et al.
Endocannabinoids may mediate the ability of (n-3) fatty acids to reduce ectopic
fat and inflammatory mediators in obese zucker rats. J Nutr 2009;139:1495501.
[84] Berger A, Gershwin ME, German JB. Effects of various dietary fats on cardiolipin
acyl composition during ontogeny of mice. Lipids 1992;27:60512.
[85] Malis CD, Weber PC, Leaf A, Bonventre JV. Incorporation of marine lipids into
mitochondrial membranes increases susceptibility to damage by calcium and
reactive oxygen species: evidence for enhanced activation of phospholipase A2 in
mitochondria enriched with n-3 fatty acids. Proc Natl Acad Sci U S A 1990;87:88459.
[86] Plourde M, Chouinard-Watkins R, Rioux-Perreault C, Fortier M, Dang MT, Allard
MJ, et al. Kinetics of 13C-DHA before and during fish-oil supplementation in
healthy older individuals. Am J Clin Nutr 2014;100:10512.
[87] Ghasemifard S, Hermon K, Turchini GM, Sinclair AJ. Metabolic fate (absorption,
beta-oxidation and deposition) of long-chain n-3 fatty acids is affected by sex and
by the oil source (krill oil or fish oil) in the rat. Br J Nutr 2015:19.
[88] Chen CT, Kitson AP, Hopperton KE, Domenichiello AF, Trepanier MO, Lin LE, et al.
Plasma non-esterified docosahexaenoic acid is the major pool supplying the brain.
Sci Rep 2015;5:15791.
[89] Sinclair AJ. Incorporation of radioactive polyunsaturated fatty acids into liver and
brain of developing rat. Lipids 1975;10:17584.
[90] Sheaff Greiner RC, Zhang Q, Goodman KJ, Giussani DA, Nathanielsz PW, Brenna JT.
Linoleate, alpha-linolenate, and docosahexaenoate recycling into saturated and
monounsaturated fatty acids is a major pathway in pregnant or lactating adults
and fetal or infant rhesus monkeys. J Lipid Res 1996;37:267586.
[91] Rhee SK, Kayani AJ, Ciszek A, Brenna JT. Desaturation and interconversion of
dietary stearic and palmitic acids in human plasma and lipoproteins. Am J Clin
Nutr 1997;65:4518.
[92] Emken EA, Rohwedder WK, Adlof RO, Rakoff H, Gulley RM. Metabolism in humans
of cis-12,trans-15-octadecadienoic acid relative to palmitic, stearic, oleic and
linoleic acids. Lipids 1987;22:495504.
[93] Wijendran V, Huang MC, Diau GY, Boehm G, Nathanielsz PW, Brenna JT. Efficacy
of dietary arachidonic acid provided as triglyceride or phospholipid as substrates
for brain arachidonic acid accretion in baboon neonates. Pediatr Res 2002;51:
26572.
[94] Igarashi M, DeMar Jr JC, Ma K, Chang L, Bell JM, Rapoport SI. Upregulated liver
conversion of alpha-linolenic acid to docosahexaenoic acid in rats on a 15 week n-
3 PUFA-deficient diet. J Lipid Res 2007;48:15264.
[95] Chowdhury R, Warnakula S, Kunutsor S, Crowe F, Ward HA, Johnson L, et al.
Association of dietary, circulating, and supplement fatty acids with coronary risk:
a systematic review and meta-analysis. Ann Intern Med 2014;160:398406.