Comparison of Low Fat and Low Carbohydrate Diets on Circulating Fatty Acid Composition and Markers of Inflammation

Abstract

Abnormal distribution of plasma fatty acids and increased inflammation are prominent features of metabolic syndrome. We tested whether these components of metabolic syndrome, like dyslipidemia and glycemia, are responsive to carbohydrate restriction. Overweight men and women with atherogenic dyslipidemia consumed ad libitum diets very low in carbohydrate (VLCKD) (1504 kcal:%CHO:fat:protein = 12:59:28) or low in fat (LFD) (1478 kcal:%CHO:fat:protein = 56:24:20) for 12 weeks. In comparison to the LFD, the VLCKD resulted in an increased proportion of serum total n-6 PUFA, mainly attributed to a marked increase in arachidonate (20:4n-6), while its biosynthetic metabolic intermediates were decreased. The n-6/n-3 and arachidonic/eicosapentaenoic acid ratio also increased sharply. Total saturated fatty acids and 16:1n-7 were consistently decreased following the VLCKD. Both diets significantly decreased the concentration of several serum inflammatory markers, but there was an overall greater anti-inflammatory effect associated with the VLCKD, as evidenced by greater decreases in TNF-alpha, IL-6, IL-8, MCP-1, E-selectin, I-CAM, and PAI-1. Increased 20:4n-6 and the ratios of 20:4n-6/20:5n-3 and n-6/n-3 are commonly viewed as pro-inflammatory, but unexpectedly were consistently inversely associated with responses in inflammatory proteins. In summary, a very low carbohydrate diet resulted in profound alterations in fatty acid composition and reduced inflammation compared to a low fat diet.

Full text

ARTICLE
Comparison of Low Fat and Low Carbohydrate Diets
on Circulating Fatty Acid Composition and Markers
of Inflammation
Cassandra E. Forsythe Æ Stephen D. Phinney Æ Maria Luz Fernandez Æ
Erin E. Quann Æ Richard J. Wood Æ Doug M. Bibus Æ William J. Kraemer Æ
Richard D. Feinman Æ Jeff S. Volek
Received: 5 July 2007 / Revised: 24 October 2007 / Accepted: 25 October 2007 / Published online: 29 November 2007
Ó AOCS 2007
Abstract Abnormal distribution of plasma fatty acids and
increased inflammation are prominent features of meta-
bolic syndrome. We tested whether these components of
metabolic syndrome, like dyslipidemia and glycemia, are
responsive to carbohydrate restriction. Overweight men
and women with atherogenic dyslipidemia consumed
ad libitum diets very low in carbohydrate (VLCKD)
(1504 kcal:%CHO:fat:protein = 12:59:28) or low in fat
(LFD) (1478 kcal:%CHO:fat:protein = 56:24:20) for
12 weeks. In comparison to the LFD, the VLCKD resulted
in an increased proportion of serum total n-6 PUFA, mainly
attributed to a marked increase in arachidonate (20:4n-6),
while its biosynthetic metabolic intermediates were
decreased. The n-6/n-3 and arachidonic/eicosapentaenoic
acid ratio also increased sharply. Total saturated fatty acids
and 16:1n-7 were consistently decreased following the
VLCKD. Both diets significantly decreased the concen-
tration of several serum inflammatory markers, but there
was an overall greater anti-inflammatory effect associated
with the VLCKD, as evidenced by greater decreases in
TNF-a, IL-6, IL-8, MCP-1, E-selectin, I-CAM, and PAI-1.
Increased 20:4n-6 and the ratios of 20:4n-6/20:5n-3 and
n-6/n-3 are commonly viewed as pro-inflammatory, but
unexpectedly were consistently inversely associated with
responses in inflammatory proteins. In summary, a very
low carbohydrate diet resulted in profound alterations in
fatty acid composition and reduced inflammation compared
to a low fat diet.
Keywords Arachidonic acid  Palmitoleic acid 
Ketogenic diet  Saturated fat  Metabolic syndrome
Abbreviations
VLCKD Very low carbohydrate ketogenic diet
LFD Low fat diet
PL Phospholipid
CE Cholesteryl ester
CVD Cardiovascular disease
RDA Recommended daily allowance
BMI Body mass index
IL Interleukin
TNF-a Tumor necrosis factor-a
VEGF Vascular endothelial growth factor
IFN-c Interferon-c
EGF Epidermal growth factor
MCP-1 Monocyte chemotactic protein-1
ICAM-1 Intracellular cellular adhesion molecule-1
VCAM-I Vascular cellular adhesion molecule-I
NF-jB Nuclear factor-kappa B
C. E. Forsythe  E. E. Quann  W. J. Kraemer  J. S. Volek (&)
Department of Kinesiology, University of Connecticut,
2095 Hillside Road, Unit 1110, Storrs, CT 06269-1110, USA
e-mail: jeff.volek@uconn.edu
S. D. Phinney
School of Medicine, University of California, Davis, CA, USA
M. L. Fernandez  R. J. Wood  J. S. Volek
Department of Nutritional Science, University of Connecticut,
Storrs, CT, USA
D. M. Bibus
University of Minnesota and Lipid Technologies, LLC,
Austin, MN, USA
R. D. Feinman
Department of Biochemistry, SUNY Downstate Medical Center,
Brooklyn, NY, USA
123
Lipids (2008) 43:65–77
DOI 10.1007/s11745-007-3132-7

CRP C-reactive protein
TAG Triglycerides
Introduction
Development of metabolic syndrome (insulin resistance
syndrome) is associated with altered composition of cir-
culating fatty acids characterized by higher saturated fatty
acids (14:0, 16:0), higher palmitoleic acid (16:1n-7, the
MUFA product derived from palmitic acid), higher diho-
mo-c-linolenic acid (20:3n-6, the precursor of arachidonic
acid), and lower levels of linoleic acid (18:2n-6) [1]. The
effect of dietary fatty acid composition on circulating fatty
acids [2] is not well understood. Two recent studies dem-
onstrated that consumption of a diet higher in saturated fat
resulted in lower circulating palmitic acid (16:0) in cho-
lesteryl ester compared to a diet low in saturated fat [3, 4],
a paradox likely explained by the level of carbohydrate [5]
whose increase is known to be associated with de novo
fatty acid synthesis [6] and decreased fatty acid oxidation.
We have previously described a comparison between a
very low carbohydrate diet (VLCKD) and a low fat diet
(LFD) in subjects with features of metabolic syndrome. A
notable finding was an inverse relationship between dietary
and plasma saturated fatty acids (SFA). The VLCKD, with
three-fold greater dietary SFA than the LFD, showed a
consistently greater reduction in plasma SFA compared to
the LFD [7].
Metabolic syndrome is generally defined by high fasting
glucose, triglycerides (TG), blood pressure and waist cir-
cumference, and low HDL cholesterol. Resistance to the
effects of insulin provides a metabolic basis for changes in
these disparate physiologic markers as well an increasing
number of associations that extend beyond the original
description of the syndrome almost 20 years ago [8]. New
features that appear to be associated with metabolic syn-
drome include disturbed circulating fatty acid composition,
perturbed lipid metabolism and increased oxidative stress
and inflammation [9]. Fatty acids contribute to overall
inflammatory balance by several mechanisms. In macro-
phages, SFA activate toll-like receptor signaling leading to
activation of nuclear factor-kappa B (NF-jB) and expres-
sion of cyclooxygenase-2 [10, 11]. NF-jB is a transcription
factor that regulates over 100 genes, many with an estab-
lished role in inflammatory responses and atherosclerosis,
and may therefore represent a crucial link between fatty
acids, metabolic syndrome and atherogenesis [12]. Ara-
chidonic acid (20:4n-6) in membranes is commonly
assumed to have a deleterious effect on overall inflamma-
tory balance because of its enzymatic conversion to
proinflammatory, proaggregative, and vasoconstrictive
eicosanoids (e.g., prostaglandin E
2
, thromboxane A
2
, leu-
kotrienes B
4
). Arachidonic acid is also capable of non-
enzymatic conversion to other proinflammatory bioactive
products (F2-isoprostanes) via interaction with molecular
oxygen. In contrast, eicosanoids derived from the 20-car-
bon n-3 PUFA, eicosapentaenoic acid (20:5n-3), have less
potent inflammatory effects [13]. A recent report showed a
marked increase in the plasma 20:4n-6/20:5n-3 ratio in
subjects consuming a VLCKD, while CRP, a marker of
constitutive inflammation, decreased slightly [14]. The
relations between inflammatory markers and arachidonic
acid metabolism are complex [15], and may be further
modified by the level of dietary carbohydrate.
Carbohydrate restriction is generally effective at ame-
liorating those physiologic markers associated with
metabolic syndrome: high fasting glucose and insulin, and
particularly the atherogenic dyslipidemia characterized by
high TG and low HDL [16–19]. The effects are presumed
to be attributed to better regulation of plasma glucose and
insulin levels and improvement in the hyperinsulinemia/
insulin resistance that are fundamental features of meta-
bolic syndrome. Here we evaluated circulating fatty acid
composition in three lipid fractions as well as a large
number of inflammatory makers and show that a VLCKD
results in profound alterations in fatty acid composition and
reduced inflammatory markers to a greater extent than a
low fat diet.
Materials and Methods
Study Design and Subjects
Details of this investigation have been described previously
[7]. In brief, 40 overweight men and women aged
18–55 year with a BMI [25 kg/m
2
participated in this
12 week randomized, controlled, dietary intervention trial
comparing a VLCKD to a LFD. All participants were
required to have atherogenic dyslipidemia defined by
moderately elevated TG (150 to 500 mg/dl) and low HDL
[\40 (men) or \50 (women) mg/dl]. The two dietary
groups were balanced for gender, age and BMI. Exclusion
criteria were any metabolic and endocrine disorders, use of
glucose-lowering, lipid-lowering or vasoactive prescrip-
tions or supplements, consumption of a VLCKD, or weight
loss [5.0 kg in the past 3 months. Habitual physical
activity was maintained throughout the study intervention
and was documented daily. Blood was drawn at baseline
and after 12 week of diet intervention in the morning after a
12 h overnight fast and a 24 h abstinence from alcohol and
strenuous exercise. All procedures were approved by the
Institutional Review Board of the University of Connecti-
cut, and all participants provided written informed consent.
66 Lipids (2008) 43:65–77
123

Dietary Intervention
Subjects received individual and personalized dietary
counseling from Registered Dietitians during the dietary
intervention. No explicit instructions were provided
regarding caloric intake for either diet to allow expression
of any non-cognitive aspects on food intake. Subjects
received weekly follow-up counseling during which body
mass was measured, compliance was assessed, and further
dietetic education was provided. Dietary intake and com-
pliance was assessed with detailed and weighed 7-day food
records at baseline, during weeks one, 6, and 12, and was
analyzed for energy and macro/micronutrient content using
NUTRITIONIST PRO
TM
(Version 1.5, First Databank Inc,
The Hearst Corporation, San Bruno, CA, USA). The
nutrient analysis program had no missing values for the
nutrients reported and the database was extensively upda-
ted with new foods and individualized recipes. All subjects
were given a multi-vitamin/mineral complex that provided
micronutrients at levels B100% of the RDA.
The main goal of the VLCKD was to restrict carbohy-
drate to a level that induced a low level of ketosis. Subjects
monitored their level of ketosis daily using urine reagent
strips. In this diet there were no restrictions on the intake of
fat from saturated and unsaturated sources or the intake of
cholesterol. Examples of foods consumed by the subjects
included unlimited amounts of beef, poultry, fish, eggs, oils
and heavy cream; moderate amounts of hard cheeses, low
carbohydrate vegetables and salad dressings; and small
amounts of nuts, nut butters and seeds. Subjects restricted
fruit and fruit juices, dairy products (with the exception of
heavy cream and hard cheese), breads, grains, pasta, cereal,
high carbohydrate vegetables, and desserts. Subjects were
instructed to avoid all low carbohydrate breads and cereal
products, and were limited to a maximum of one sugar
alcohol-containing, low carbohydrate snack per day.
The LFD was designed to provide \10% of total calo-
ries from saturated fat and \300 mg cholesterol. Foods
encouraged included whole grains (breads, cereals, and
pastas), fruit/fruit juices, vegetables, vegetable oils, low-fat
dairy and lean meat products. Standard diabetic exchange
lists were used to ensure a macronutrient balance of protein
(*20% energy), fat (*25% energy), and carbohydrate
(*55% of energy).
Blood Analyses
Whole blood was collected into tubes with no preservative
or EDTA and centrifuged at 15009g for 15 min and 4°C,
and promptly aliquoted into separate storage tubes which
were stored at 75 °C until analyzed for serum fatty acid
composition and plasma inflammatory markers. An aliquot
of anti-coagulated whole blood (*3 ml) was sent to a
certified medical laboratory (Quest Diagnostics, Walling-
ford, CT, USA) for a white blood cell differential count.
Inflammatory Biomarkers
The Evidence Investigator
TM
Biochip Array technology
(Randox Laboratories Ltd, UK) that uses sandwich chemi-
luminescent immunoassays to simultaneously detect
multiple analytes from a single sample was used to determine
the following serum cytokines and adhesion molecules:
IL-6, IL-8, vascular endothelial growth factor (VEGF), TNF-
a, IFN-c, epidermal growth factor (EGF), monocyte che-
motactic protein-1 (MCP-1), intracellular cellular adhesion
molecule-1 (ICAM-1), vascular cellular adhesion molecule-
I (VCAM-I), E-selectin, P-selectin and L-selectin. In addi-
tion, serum C-reactive protein (CRP) was determined on an
IMMULITE Automated Analyzer using the commercially
available immulite chemiluminescent enzyme immuno-
metric assay (Immulite
1
, Diagnostic Products Corp, Los
Angeles, CA, USA) and plasma plasminogen-activator
inhibitor-1 (PAI-1) was determined utilizing the Luminex
200 analyzer (Luminex Corp, Austin, TX, USA) using an
immunoassay kit from LINCO Research.
Fatty Acid Composition
Serum lipids were extracted according to the method of
Bligh–Dyer whereby mixtures of plasma, methanol, chlo-
roform and water were prepared such that lipid is recovered
in a chloroform layer. The resulting lipid extracts were
maintained under an atmosphere of nitrogen following
extraction and kept frozen prior to additional processing.
Immediately prior to lipid class separation, lipid samples
were dried under a gentle stream of nitrogen, rediluted in
50 ll of chloroform and prepared for lipid class separation.
Lipid classes including total TAG, PL and CE were sepa-
rated on commercial silica gel G plates (AnalTech,
Newark, DE, USA). The chromatographic plates were
developed in a solvent system consisting of distilled
petroleum ether (bp 30–60 °C):diethyl ether:acetic acid
(80:20:1, by vol). Following development, the silica gel
plates were sprayed with a methanolic solution containing
0.5% 2,7-dichlorofluorescein which was then used to
visualize lipid classes under ultraviolet light. Desired cor-
responding lipid bands were then scraped into Teflon line
screw cap tubes. The samples were then transesterified with
boron trifluoride (10%) in excess methanol (Supelco,
Bellefonte, PA, USA) in an 80 °C water bath for 90 min.
Resulting fatty acid methyl esters were extracted with
water and petroleum ether and stored frozen until gas
chromatographic analysis was performed.
Lipids (2008) 43:65–77 67
123

Lipid class fatty acid methyl ester composition was
determined by capillary gas chromatography. Methyl ester
samples were blown to dryness under nitrogen and resus-
pended in hexane. Resulting fatty acid methyl esters were
separated and quantified with a Shimadzu capillary gas
chromatograph (GC17) utilizing a 30 m Restek free fatty
acid phase (FFAP) coating and EZChrom software. The
instrument temperature was programmed from 190 to
240 °Cat7°C/min with a final hold of 10 min, separating
and measuring fatty acid methyl esters ranging from 12:0 to
24:1. The detector temperature was 250 °C. Helium carrier
gas was used at a flow rate of 1.4 ml/min. and a split ratio of
1:25. Chromatographic data was collected and processed
with EZChrom software (Scientific Products, CA, USA).
Fatty acids were identified by comparison to authentic fatty
acid standards and quantitated with peak area and internal
standard. Resulting data are expressed in percent compo-
sition. Individual peaks, representing as little as 0.05% of
the fatty acid methyl esters, were distinguished.
Statistical Analyses
All statistical analyses were done with Statistica software
(StatSoft Inc, Tulsa, OK, USA). A 2 9 2 ANOVA with
one between effect (VLCKD vs. LFD) and one within
effect (Week 0 vs. Week 12) was used to compare bio-
chemical responses over time in both groups. Significant
main or interaction effects were further analyzed using a
Fishers LSD post hoc test. Relationships among selected
variables were examined using Pearson’s product-moment
correlation coefficient. The alpha level for significance was
set at 0.05.
Results
Dietary Intake and MetS Responses
Dietary nutrient intake and responses of MetS biomarkers
will be presented elsewhere [7]. In brief, subjects in both
groups reduced energy intake to approximately 1500 kcal/
day, but the diets had markedly different macronutrient
distributions based upon the analysis of individual diet
records (VLCKD, %CHO:fat:protein = 12:59:28) and
(LFD, %CHO:fat:protein = 56:24:20) (Table 1). Dietary
saturated fat and cholesterol intake were significantly
higher during the VLCKD than the LFD. The LFD led to
improvements in some metabolic markers, but subjects
following the VLCKD had consistently greater weight loss,
Table 1 Daily nutrient intake and serum cholesterol responses of men and women who consumed low carbohydrate and low fat diets
Variables VLCKD (n = 20) LFD (n = 20) 2 9 2 ANOVA
Baseline Intervention Baseline Intervention Time T 9 G
Energy (kcal) 2351 ± 617 1504 ± 494 2082 ± 445 1478 ± 435 0.000 0.154
Protein (g) 94.6 ± 28.5 104.8 ± 33.6 82.3 ± 17.6 71.5 ± 21.3 0.756 0.009
Protein (%) 16.2 ± 3.1 28.1 ± 4.4 15.8 ± 2.6 19.6 ± 4.4 0.000 0.000
Carbohydrate (g) 270.3 ± 67.2 44.8 ± 18.9 266.8 ± 74.7 208.3 ± 69.6 0.000 0.000
Carbohydrate (%) 46.6 ± 7.7 12.4 ± 5.2 50.9 ± 10.1 55.8 ± 7.9 0.000 0.000
Total Fat (g) 97.0 ± 35.2 100.2 ± 37.9 78.5 ± 29.5 40.0 ± 17.5 0.004 0.001
Total Fat (%) 36.2 ± 6.7 58.9 ± 5.4 33.0 ± 9.8 23.8 ± 6.8 0.000 0.000
Saturated Fat (g) 34.2 ± 14.3 36.4 ± 12.9 26.0 ± 11.1 11.7 ± 5.9 0.012 0.002
Monounsaturated Fat (g) 19.2 ± 6.5 26.4 ± 11.1 18.0 ± 9.6 8.9 ± 4.7 0.830 0.000
Polyunsaturated Fat (g) 12.4 ± 7.4 12.4 ± 7.8 10.6 ± 7.4 5.1 ±
3.1 0.064 0.019
18:1n-9 (g) 14.0 ± 5.0 20.9 ± 9.5 11.6 ± 6.9 6.5 ± 3.8 0.289 0.000
18:2n-6 (g) 7.4 ± 6.0 7.7 ± 5.1 5.8 ± 4.7 2.9 ± 2.0 0.215 0.042
18:3n-3 (mg) 989 ± 1199 879 ± 746 575 ± 398 325 ± 198 0.139 0.439
20:5n-3 (mg) 8 ± 10 46 ± 81 32 ± 50 32 ± 58 0.047 0.050
22:6n-3 (mg) 24 ± 24 117 ± 184 83 ± 116 82 ± 154 0.049 0.052
Alcohol (%) 0.9 ± 1.8 0.7 ± 1.4 0.3 ± 0.5 0.9 ± 2.0 0.232 0.056
Cholesterol (mg) 354 ± 120 605 ± 262 267 ± 111 144 ± 80 0.044 0.000
Dietary Fiber (g) 13.1 ± 3.5 9.4 ± 4.9 15.8 ± 6.6 17.3 ± 9.6 0.083 0.021
Serum total cholesterol (mg/dl) 208.0 ± 26.0 196.5 ± 34.9 204.0 ± 31.5 194.5 ± 34.0 0.016 0.816
Serum LDL-C (mg/dL) 130.4 ± 21.8 135.4 ± 31.4 127.9 ± 31.3 125.9 ± 32.1 0.357 0.357
Serum HDL-C (mg/dl) 35.8 ± 6.9 40.4 ± 9.6 38.7 ± 6.2 38.4 ± 5.5 0.001 0.000
Values are mean
± SD calculated from 7 days of weight food records at baseline (week 0) and 7 days during weeks 1, 6, and 12 (Intervention)
68 Lipids (2008) 43:65–77
123

decreased adiposity, improved glycemic control and insulin
sensitivity and more favorable TAG, HDL-C and total
cholesterol/HDL-C ratio responses. In addition to these
markers for MetS, the VLCKD subjects showed more
favorable responses in alternative indicators of atherogenic
dyslipidemia and cardiovascular risk: postprandial lipemia,
apo B, apo A-1, the apo B/Apo A-1 ratio, LDL particle
distribution and postabsorptive and postprandial vascular
function. Most striking, we reported that despite a three-
fold higher intake of dietary saturated fat during the
VLCKD compared to the LFD, circulating saturated fatty
acids in TAG and CE were significantly decreased, as was
16:1n-7, an endogenous marker of lipogenesis. There were
profound changes, as well, in other fatty acids in circulat-
ing TG, PL, and CE fractions (Tables 2–4).
Circulating Triglyceride Fatty Acids
Compared to the LFD, consumption of the VLCKD
resulted in a significantly greater increase in TG n-6 PUFA,
mainly attributed to a marked increase in arachidonic acid:
17 of 20 subjects in VLCKD showed marked increases
while only 7 of 20 subjects on LFD showed increases in
20:4n-6 and these were more modest in amplitude
(Table 2). In both groups, n-3 PUFA was decreased due
largely to lower a-linolenic acid (18:3n-3) and 20:5n-3.
The n-6/n-3 and arachidonic/eicosapentaenoic acid ratios
were, on average, nearly doubled in response to the
VLCKD and virtually unchanged by the LFD: 15 of 20
subjects on VLCKD showed increases in the n-6/n-3 ratios,
while only 8 of 20 of the LFD showed increases. In contrast
to the response in 20:4n-6, the metabolic intermediates in
the biosynthetic pathway, especially 20:3n-6, were
decreased after the VLCKD. As previously reported, total
MUFA was unchanged but 16:1n-7 and total SFA was
significantly decreased in response to the VLCKD.
Circulating Phospholipid Fatty Acids
The pattern of fatty acid changes seen in the TAG fraction
was also found in circulating phospholipids: consumption
of the VLCKD was associated with an increase in n-6
PUFA, again primarily due to a distinct increase in 20:4n-6
(Fig 1a), whereas 18:3n-6 and 20:3n-6 were markedly
decreased (Table 3). The VLCKD was associated with a
significant reduction in 18:3n-3, 18:4n-3, 20:4n-3, and
20:5n-3. However docosahexaenoic acid (22:6n-3) was
increased so that total n-3 PUFA was not significantly
changed. Compared to the LFD, ingestion of the VLCKD
resulted in a significant increase in total PUFA and the ratio
of n-6/n-3 (Fig. 1b) and arachidonic/eicosapentaenoic acid
(Fig. 1c). In comparison to the LFD, total MUFA was
significantly decreased in response to the VLCKD due to
significant decreases in the most abundant MUFA, 18:1n-9,
and a consistent reduction in 16:1n-7.
Circulating Cholesteryl Ester Fatty Acids
The patterns of CE fatty acid responses to diet resembled
that in TG and PL, and in general were more dramatic in
this lipid fraction (Table 4). Total n-6 PUFA was signifi-
cantly increased in response to the VLCKD due to a large
increase in 18:2n-6 and a smaller increase in 20:4n-6: most
of the VLCKD group showed a significant increase in
20:4n-6 while only five of the LFD group showed an
increase. On the other hand total n-3 PUFA was signifi-
cantly decreased due to a reduced proportion of 18:3n-3
and 20:5n-3 in response to the VLCKD but, similar to other
fractions, there was an increase in 22:6n-3. The n-6/n-3 and
arachidonic/eicosapentaenoic acid ratios were unchanged
in response to the LFD but increased sharply after con-
sumption of the VLCKD. Total MUFA decreased in
response to the VLCKD, again due to a reduced proportion
of 18:1n-9 and a striking decrease in 16:1n-7.
Inflammatory Markers
Both diets led to a similar significant reduction in the acute
phase reactant C-reactive protein (-23%), VEGF (-21%),
P-selectin (-11%), and a trend for EGF (-38%), and
V-CAM (-6%); however, there was an overall greater
anti-inflammatory effect associated with the VLCKD as
evidenced by significantly greater decreases in the proin-
flammatory cytokine TNF-a (-32 vs. -12%), the
chemokines IL-8 (-33 vs. 4%) and MCP-1 (-24 vs.
-5%), and the adhesion molecules E-selectin (
-34 vs.
-14%) and I-CAM (-17 vs. -3%) (Table 5). There was
also a trend for a greater reduction in IL-6 (-35%) in
response to the VLCKD (P = 0.07). Plasminogen-activator
inhibitor-1 (PAI-1) has antifibrinolytic functions, and was
also reduced more in subjects consuming the VLCKD
compared to LFD (-34 vs. -8%). There was no effect of
the interventions on leukocyte subpopulations.
Correlations
The results described above are surprising in that con-
sumption of the VLCKD showed substantially greater
increases in arachidonic acid and the arachidonic/eicosa-
pentaenoic acid and n-6/n-3 ratios that are commonly
viewed as contributing to an overall proinflammatory state,
Lipids (2008) 43:65–77 69
123

Table 2 Serum triglyceride fatty acid responses of men and women who consumed low carbohydrate and low fat diets
Variables VLCKD (n = 20) Low Fat (n = 20) 2 9 2 ANOVA
Week 0 Week 12 Week 0 Week 12 Time T 9 G
Total TG (mg/dl) 210.9 ± 57.9 103.7 ± 44.1 187.1 ± 57.6 151.2 ± 38.0 0.000 0.000
SFA
12:0 0.09 ± 0.09 0.04 ± 0.06 0.06 ± 0.10 0.06 ± 0.13 0.271 0.108
14:0 1.94 ± 0.61 1.02 ± 0.40 1.86 ± 0.88 1.71 ± 0.61 0.000 0.001
15:0 0.27 ± 0.07 0.25 ± 0.05 0.28 ± 0.10 0.28 ± 0.08 0.478 0.527
16:0 27.07 ± 4.14 24.26 ± 1.65 24.86 ± 2.31 23.99 ± 1.86 0.002 0.092
18:0 3.66 ± 1.06 3.45 ± 0.66 3.32 ± 1.09 2.86 ± 0.48 0.051 0.466
20:0 0.06 ± 0.03 0.05 ± 0.04 0.04 ± 0.03 0.04 ± 0.03 0.354 0.901
22:0 0.03 ± 0.03 0.05 ± 0.03 0.03 ± 0.03 0.04 ± 0.03 0.079 0.584
24:0 0.01 ± 0.02 0.02 ± 0.04 0.00 ± 0.01 0.00 ± 0.00 0.514 0.155
Total SFA 33.13 ± 5.03 29.15 ± 1.39 30.45
± 3.98 28.98 ± 2.45 0.000 0.086
MUFA
14:1 0.22 ± 0.15 0.06 ± 0.09 0.18 ± 0.17 0.17 ± 0.13 0.000 0.004
15:1 0.04 ± 0.03 0.04 ± 0.05 0.03 ± 0.02 0.04 ± 0.04 0.627 0.659
16:1n-7 4.53 ± 1.07 3.10 ± 0.69 4.54 ± 0.98 4.53 ± 1.10 0.000 0.000
17:1 0.29 ± 0.06 0.26 ± 0.08 0.29 ± 0.10 0.29 ± 0.07 0.248 0.264
18:1n-9 36.31 ± 3.18 38.67 ± 2.73 38.91 ± 2.75 39.50 ± 2.90 0.015 0.132
20:1n-7 0.07 ± 0.03 0.05 ± 0.02 0.07 ± 0.02 0.07 ± 0.03 0.015 0.100
20:1n-9 0.27 ± 0.07 0.27 ± 0.08 0.31 ± 0.07 0.31 ± 0.07 0.912 0.672
22:1n-9 0.05 ± 0.03 0.05 ± 0.03 0.04 ± 0.03 0.03 ± 0.03 0.454 0.352
Total MUFA 41.78 ± 3.51 42.50 ± 2.99 44.37 ± 2.92 44.93 ± 3.25 0.292 0.900
n-6 PUFA
18:2n-6 19.04 ± 3.58 22.19 ± 3.09 19.36 ± 4.36 19.99 ± 3.09 0.010 0.078
18:3n-6 0.45 ± 0.19 0.38 ± 0.17 0.38
± 0.18 0.41 ± 0.23 0.517 0.102
20:2n-6 0.20 ± 0.07 0.18 ± 0.07 0.21 ± 0.08 0.21 ± 0.09 0.616 0.259
20:3n-6 0.34 ± 0.10 0.26 ± 0.06 0.33 ± 0.07 0.33 ± 0.08 0.001 0.003
20:4n-6 1.17 ± 0.35 1.70 ± 0.55 1.04 ± 0.27 1.16 ± 0.35 0.000 0.002
22:4n-6 0.18 ± 0.04 0.21 ± 0.06 0.17 ± 0.04 0.17 ± 0.04 0.040 0.049
22:5n-6 0.14 ± 0.07 0.27 ± 0.18 0.15 ± 0.10 0.13 ± 0.07 0.033 0.004
Total n-6 21.51 ± 3.59 25.18 ± 2.86 21.64 ± 4.24 22.39 ± 3.13 0.004 0.048
n-3 PUFA
18:3n-3 1.10 ± 0.44 0.90 ± 0.51 1.04 ± 0.35 1.13 ± 0.60 0.494 0.080
18:4n-3 0.39 ± 0.21 0.27 ± 0.17 0.46 ± 0.20 0.44 ± 0.18 0.009 0.042
20:3n-3 0.04 ± 0.03 0.02 ± 0.02 0.03 ± 0.02 0.03 ± 0.04 0.054 0.051
20:4n-3 0.06 ± 0.04 0.02 ± 0.03 0.04 ± 0.03 0.04 ± 0.04 0.019 0.005
20:5n-3 0.15 ± 0.05 0.12 ± 0.04 0.13 ±
0.06 0.16 ± 0.11 0.696 0.025
22:5n-3 0.24 ± 0.06 0.29 ± 0.09 0.25 ± 0.08 0.28 ± 0.10 0.011 0.497
22:6n-3 0.31 ± 0.16 0.42 ± 0.16 0.27 ± 0.11 0.33 ± 0.23 0.019 0.486
Total n-3 2.30 ± 0.57 2.05 ± 0.63 2.22 ± 0.44 2.41 ± 0.72 0.790 0.039
n-9 PUFA
20:3n-9 0.09 ± 0.04 0.12 ± 0.05 0.12 ± 0.06 0.13 ± 0.09 0.012 0.455
Total PUFA 23.90 ± 3.91 27.35 ± 3.11 23.97 ± 4.34 24.94 ± 3.56 0.006 0.106
Total HUFA 4.66 ± 0.73 4.98 ± 0.75 4.41 ± 0.64 4.74 ± 0.97 0.034 0.986
Total n-9 36.85 ± 3.24 39.05 ± 2.77 39.44 ± 2.68 40.02 ± 2.84 0.023 0.170
n-6/n-3 9.74 ± 2.20 13.16 ± 3.49 10.04 ± 2.47 9.74 ± 2.26 0.002 0.000
20:4n-6/20:5n-3 7.69 ± 1.82 14.76 ± 6.21 8.73 ± 3.44 9.33 ± 6.05 0.000 0.002
Values are weight percent (mean ± SD)
70 Lipids (2008) 43:65–77
123

while simultaneously there was a significant decrease in
many inflammatory markers. An analysis of these data
bears out the idea that changes in the fatty acid proxies
were consistently inversely associated with responses in
most of the inflammatory markers we measured (Fig. 2).
The fatty acid with the most consistent positive association
with changes in inflammatory markers was palmitoleic acid
(16:1n-7). The correlations between weight loss and
changes in inflammatory markers were generally small and
not significant. As shown in Fig. 3 for two of the more
important markers TNF-a and IL-8, there is essentially no
correlation.
Discussion
Because of the continued emphasis on dietary recommen-
dations for cardiovascular disease and general health, the
relation between dietary fat intake and plasma fatty acids
and inflammatory markers is of great importance. The
findings presented here support our hypothesis that the
components of metabolic syndrome are distinctly those that
respond favorably to reduction in dietary carbohydrate
[19]. Responses in fatty acid composition to the VLCKD in
this study were exactly opposite to the fatty acid profile
recently shown to be associated with development of
metabolic syndrome over a 20 year period in previously
healthy men (i.e., higher circulating 14:0, 16:0, 16:1n-7,
18:1n-9, 18:3n-6, and 20:3n-6, and lower levels of 18:2n-6)
[1]. Abnormal fatty acid composition and inflammatory
status are now recognized as prominent features of meta-
bolic syndrome, and are reliably improved in subjects
consuming a low carbohydrate diet compared to a low fat
diet.
Acute ingestion of carbohydrate clearly induces an
increase in reactive oxygen species and activation of pro-
inflammatory pathways [9], and isocaloric high carbohy-
drate [20] and high glycemic [21] diets are associated with
increased biomarkers of inflammation. In the context of
hypocaloric diets, we showed that reducing dietary total
6
8
01
21
4
1
6
1
8
1
0
2
21
k
W
0kW
20:4n-6 in PL
DKC
L
V
A
6
8
01
2
1
41
61
81
02
21kW0
kW
taFwoL
2
4
6
8
01
2
1
41
61
21k
W
0
kW
n-6/n-3 in PL
D
K
C
L
V
B
0
01
02
0
3
04
05
0
6
07
08
0
9
21kW0k
W
t
aF
wo
L
0
01
02
03
04
0
5
06
07
08
0
9
21kW0kW
20:4n-6/20:5n-3 in PL
DK
CLV
C
2
4
6
8
0
1
2
1
4
1
61
2
1k
W0kW
t
a
F
woL
Fig. 1 Individual responses in serum phospholipid arachidonic acid
(20:4n-6) (a), the n-6/n-3 ratio (b) and the arachidonic/eicosapenta-
enoic acid ratio (20:4n-6/20:5n-3) (c) in subjects who consumed a
very low carbohydrate ketogenic diet (VLCKD) or a low fat diet for
12 weeks. Shaded bars indicate mean responses
Lipids (2008) 43:65–77 71
123

Table 3 Serum phospholipid fatty acid responses of men and women who consumed low carbohydrate and low fat diets
Variables VLCKD (n = 20) Low fat (n = 20) 2 9 2 ANOVA
Week 0 Week 12 Week 0 Week 12 Time T 9 G
Total phospholipids (mg/dl) 196 ± 25 170 ± 33 183 ± 28 173 ± 30 0.001 0.135
SFA
14:0 0.34 ± 0.12 0.28 ± 0.12 0.35 ± 0.13 0.36 ± 0.15 0.186 0.094
15:0 0.16 ± 0.03 0.17 ± 0.03 0.18 ± 0.06 0.18 ± 0.05 0.224 0.681
16:0 26.62 ± 1.64 26.87 ± 1.52 26.46 ± 1.59 27.02 ± 1.49 0.107 0.512
18:0 13.83 ± 1.25 12.76 ± 0.97 14.20 ± 1.58 13.56 ± 1.22 0.000 0.217
20:0 0.13 ± 0.04 0.18 ± 0.04 0.13 ± 0.03 0.13 ± 0.03 0.000 0.001
22:0 0.41 ± 0.11 0.57 ± 0.21 0.36 ± 0.11 0.35 ± 0.07 0.000 0.000
24:0 0.35 ± 0.11 0.48 ± 0.21 0.27 ± 0.12 0.26 ± 0.10 0.006 0.001
Total SFA 41.84 ± 1.30 41.31 ± 1.45 41.94 ± 1.17 41.88 ± 1.00 0.127 0.231
MUFA
14:1 0.03 ± 0.06 0.07 ± 0.09 0.09 ±
0.09 0.10 ± 0.09 0.134 0.331
15:1 0.35 ± 0.11 0.54 ± 0.21 0.25 ± 0.16 0.29 ± 0.14 0.001 0.035
16:1n-7 0.92 ± 0.23 0.60 ± 0.17 0.86 ± 0.18 0.86 ± 0.23 0.000 0.000
17:1 0.28 ± 0.10 0.41 ± 0.17 0.26 ± 0.10 0.27 ± 0.08 0.002 0.007
18:1n-9 10.95 ± 1.60 9.88 ± 1.17 11.38 ± 0.94 11.67 ± 1.26 0.122 0.009
20:1n-7 0.05 ± 0.03 0.03 ± 0.01 0.04 ± 0.02 0.04 ± 0.02 0.011 0.020
20:1n-9 0.10 ± 0.02 0.12 ± 0.03 0.12 ± 0.05 0.13 ± 0.03 0.000 0.244
22:1n-9 0.05 ± 0.06 0.06 ± 0.04 0.06 ± 0.06 0.06 ± 0.05 0.754 0.654
24:1 0.42 ± 0.13 0.58 ± 0.22 0.37 ± 0.11 0.42 ± 0.11 0.000 0.041
Total MUFA 13.14 ± 1.68 12.29 ± 1.14 13.43 ± 1.09 13.84 ± 1.37 0.392 0.015
n-6 PUFA
18:2n-6 22.99 ± 3.09 23.74 ± 2.69 22.66 ± 3.09 22.13 ± 2.74 0.807 0.162
18:3n-6 0.17 ± 0.06 0.10 ± 0.02 0.14 ± 0.04 0.14
± 0.03 0.000 0.002
20:2n-6 0.29 ± 0.05 0.25 ± 0.07 0.34 ± 0.09 0.32 ± 0.08 0.008 0.234
20:3n-6 3.67 ± 0.82 2.29 ± 0.55 3.78 ± 0.76 3.50 ± 0.72 0.000 0.000
20:4n-6 11.54 ± 2.05 13.70 ± 1.82 11.31 ± 2.05 11.48 ± 1.97 0.000 0.001
22:4n-6 0.44 ± 0.08 0.40 ± 0.09 0.43 ± 0.10 0.42 ± 0.09 0.023 0.222
22:5n-6 0.41 ± 0.13 0.39 ± 0.11 0.38 ± 0.10 0.36 ± 0.13 0.379 0.883
Total n-6 39.50 ± 2.17 40.86 ± 2.04 39.04 ± 1.53 38.34 ± 1.53 0.246 0.001
n-3 PUFA
18:3n-3 0.20 ± 0.08 0.15 ± 0.06 0.15 ± 0.07 0.19 ± 0.13 0.684 0.015
18:4n-3 0.13 ± 0.04 0.08 ± 0.04 0.13 ± 0.05 0.13 ± 0.07 0.000 0.003
20:3n-3 0.03 ± 0.02 0.03 ± 0.02 0.04 ± 0.03 0.04 ± 0.04 0.752 0.999
20:4n-3 0.12 ± 0.06 0.03 ± 0.02 0.10 ± 0.02 0.09 ± 0.05 0.000 0.000
20:5n-3 0.61 ± 0.20 0.42 ± 0.15 0.60 ± 0.27 0.68 ±
0.48 0.264 0.009
22:5n-3 0.81 ± 0.15 0.74 ± 0.16 0.87 ± 0.22 0.86 ± 0.20 0.044 0.104
22:6n-3 2.43 ± 0.84 2.80 ± 0.81 2.48 ± 0.62 2.71 ± 0.83 0.023 0.592
Total n-3 4.33 ± 0.90 4.24 ± 0.89 4.37 ± 0.80 4.69 ± 1.19 0.419 0.164
n-9 PUFA
20:3n-9 0.09 ± 0.03 0.06 ± 0.03 0.10 ± 0.06 0.10 ± 0.05 0.048 0.092
Total PUFA 43.92 ± 1.89 45.17 ± 1.76 43.51 ± 1.11 43.14 ± 1.16 0.073 0.001
Total HUFA 20.64 ± 2.58 21.19 ± 1.90 20.51 ± 2.50 20.68 ± 2.32 0.323 0.599
Total n-9 11.55 ± 1.59 10.71 ± 1.16 12.02 ± 0.98 12.38 ± 1.29 0.341 0.021
n-6/n-3 9.55 ± 2.29 10.11 ± 2.46 9.24 ± 1.75 8.58 ± 1.84 0.879 0.062
20:4n-6/20:5n-3 21.59 ± 11.05 37.41 ± 16.28 21.86 ± 8.52 20.52 ± 8.20 0.000 0.000
Values are weight percent (mean ± SD)
72 Lipids (2008) 43:65–77
123

Table 4 Serum cholesteryl ester fatty acid responses of men and women who consumed low carbohydrate and low fat diets
Variables VLCKD (n = 20) Low fat (n = 20) 2 9 2 ANOVA
Week 0 Week 12 Week 0 Week 12 Time T 9 G
Total cholesteryl ester (mg/dl) 224 ± 35 224 ± 47 215 ± 42 199 ± 31 0.337 0.308
SFA
14:0 0.79 ± 0.18 0.51 ± 0.13 0.69 ± 0.23 0.59 ± 0.24 0.000 0.010
15:0 0.17 ± 0.07 0.15 ± 0.05 0.14 ± 0.07 0.13 ± 0.09 0.274 0.475
16:0 11.23 ± 1.13 10.40 ± 0.84 10.79 ± 1.03 10.67 ± 0.88 0.011 0.052
18:0 1.15 ± 0.28 1.00 ± 0.22 1.29 ± 0.64 1.24 ± 0.69 0.268 0.596
20:0 0.05 ± 0.05 0.06 ± 0.03 0.05 ± 0.04 0.09 ± 0.09 0.009 0.035
22:0 0.01 ± 0.03 0.01 ± 0.01 0.01 ± 0.01 0.00 ± 0.01 0.222 0.501
24:0 0.01 ± 0.01 0.00 ± 0.01 0.02 ± 0.09 0.01 ± 0.05 0.737 0.863
Total SFA 13.41 ± 1.48 12.13 ± 0.91 12.98 ± 1.29 12.74 ± 1.21 0.002 0.028
MUFA
14:1 0.79 ± 0.59 0.71 ± 0.56 0.94 ±
0.66 0.94 ± 0.63 0.785 0.779
15:1 0.05 ± 0.03 0.04 ± 0.05 0.02 ± 0.02 0.04 ± 0.09 0.606 0.218
16:1n-7 3.28 ± 0.90 1.84 ± 0.46 3.02 ± 1.01 2.98 ± 1.20 0.000 0.000
17:1 0.17 ± 0.09 0.14 ± 0.08 0.18 ± 0.08 0.16 ± 0.08 0.154 0.829
18:1n-9 17.77 ± 2.89 16.47 ± 1.69 17.52 ± 1.77 17.97 ± 1.44 0.290 0.034
20:1n-9 0.06 ± 0.06 0.04 ± 0.03 0.10 ± 0.10 0.11 ± 0.12 0.653 0.499
22:1n-9 0.02 ± 0.05 0.01 ± 0.03 0.04 ± 0.16 0.01 ± 0.03 0.260 0.583
24:1 0.00 ± 0.00 0.00 ± 0.02 0.03 ± 0.12 0.00 ± 0.00 0.444 0.263
Total MUFA 21.95 ± 2.47 22.22 ± 1.71 22.14 ± 3.65 19.26 ± 2.06 0.015 0.004
n-6 PUFA
18:2n-6 52.41 ± 4.90 56.44 ± 4.08 53.23 ± 4.52 53.05 ± 3.60 0.014 0.008
18:3n-6 1.05 ± 0.40 0.50 ± 0.16 0.95 ± 0.33 0.92 ± 0.28 0.000 0.000
20:2n-6 0.07 ± 0.10 0.04 ± 0.03 0.05 ± 0.06 0.04
± 0.05 0.093 0.483
20:3n-6 0.90 ± 0.23 0.59 ± 0.13 0.89 ± 0.17 0.85 ± 0.17 0.000 0.000
20:4n-6 7.39 ± 1.73 7.78 ± 1.73 7.57 ± 1.73 7.74 ± 1.75 0.001 0.008
22:4n-6 0.00 ± 0.00 0.00 ± 0.01 0.00 ± 0.01 0.00 ± 0.00 0.672 0.173
22:5n-6 0.05 ± 0.03 0.04 ± 0.04 0.09 ± 0.18 0.08 ± 0.16 0.855 0.997
Total n-6 61.87 ± 5.22 66.57 ± 2.89 62.79 ± 3.23 62.69 ± 2.47 0.003 0.002
n-3 PUFA
18:3n-3 0.60 ± 0.15 0.45 ± 0.14 0.51 ± 0.11 0.53 ± 0.21 0.056 0.009
18:4n-3 0.14 ± 0.05 0.08 ± 0.04 0.14 ± 0.04 0.10 ± 0.09 0.001 0.363
20:3n-3 0.03 ± 0.03 0.03 ± 0.05 0.04 ± 0.07 0.02 ± 0.03 0.169 0.320
20:4n-3 0.05 ± 0.07 0.01 ± 0.03 0.03 ± 0.04 0.02 ± 0.03 0.002 0.128
20:5n-3 0.50 ± 0.16 0.37 ± 0.10 0.57 ± 0.26 0.61 ± 0.37 0.154 0.017
22:5n-3 0.03 ± 0.08 0.01 ± 0.01 0.02 ± 0.02 0.03 ±
0.06 0.759 0.165
22:6n-3 0.40 ± 0.12 0.45 ± 0.14 0.37 ± 0.12 0.44 ± 0.18 0.012 0.520
Total n-3 1.74 ± 0.27 1.39 ± 0.23 1.67 ± 0.33 1.75 ± 0.52 0.015 0.000
Total PUFA 63.61 ± 5.15 67.96 ± 2.85 64.46 ± 3.15 64.44 ± 2.37 0.004 0.004
Total HUFA 1.13 ± 1.58 11.48 ± 1.88 11.17 ± 1.98 11.35 ± 1.84 0.308 0.743
Total n-9 18.64 ± 3.18 17.25 ± 1.78 18.63 ± 1.83 19.02 ± 1.37 0.280 0.057
n-6/n-3 36.69 ± 8.46 49.59 ± 10.34 39.24 ± 9.32 38.26 ± 9.35 0.000 0.000
20:4n-6/20:5n-3 16.24 ± 6.80 26.56 ± 11.83 15.63 ± 7.87 14.59 ± 4.94 0.001 0.000
Values are weight percent (mean ± SD)
Lipids (2008) 43:65–77 73
123

and saturated fat only had a small effect on circulating
inflammatory markers whereas reducing carbohydrate led
to considerably greater reductions in a number of
proinflammatory cytokines, chemokines, and adhesion
molecules. These data implicate dietary carbohydrate
rather than dietary fat as a more significant nutritional
factor contributing to inflammatory processes; although
increased fat in the presence of high carbohydrate may be
particularly deleterious. Dietary carbohydrate also has a
fundamental role in determining fatty acid composition of
lipids and membranes, and it is the endogenous fatty acids
(as opposed to the exogenous dietary fatty acids) that
influence inflammation by acting as ligands for receptors or
transcription factors that regulate inflammatory signaling
cascades or serving as substrates for proinflammatory
bioactive products.
Table 5 Inflammatory responses of men and women who consumed low carbohydrate and low fat diets
Variables VLCKD (n = 20) Low fat (n = 20) 2 9 2 ANOVA
Week 0 Week 12 Week 0 Week 12 Time T 9 G
WBC (910
9
/l) 6.2 ± 1.4 5.9 ± 1.4 5.9 ± 1.8 5.9 ± 2.2 0.471 0.427
Neutrophils (910
9
/ll) 3507 ± 877 3460 ± 1172 3368 ± 1500 3428 ± 1460 0.971 0.783
Lymphocytes (910
9
/ll) 2039 ± 648 1741 ± 439 1899 ± 509 1892 ± 668 0.076 0.090
Monocytes (910
9
/ll) 402 ± 135 451 ± 215 506 ± 274 422 ± 153 0.659 0.106
Eosinophils (910
9
/ll) 210 ± 169 176 ± 130 184 ± 105 164 ± 81 0.297 0.785
Basophils (910
9
/ll) 26 ± 21 21 ± 17 24 ± 13 27 ± 17 0.752 0.310
CRP (mg/dl) 0.6 ± 0.6 0.5 ± 0.5 0.4 ± 0.5 0.3 ± 0.4 0.028 0.858
IL-6 (pg/ml) 8.4 ± 9.3 5.5 ± 7.6 6.3 ± 8.7 6.3 ± 9.0 0.064 0.073
IL-8 (pg/ml) 8.5 ± 4.4 5.7 ± 2.8 9.4 ± 9.7 9.8 ± 11.0 0.033 0.007
VEGF (pg/ml) 162 ± 130 122 ± 96 129 ± 130 109 ± 113 0.004 0.293
TNF-a (pg/ml) 2.8 ± 1.6 1.9 ± 2.0 2.6 ± 1.8 2.3 ± 1.7 0.000 0.017
IFN-c (pg/ml) 2.2 ± 1.5 1.7 ± 1.9 2.0 ± 1.4 2.2 ± 1.5 0.543 0.261
EGF (pg/ml) 12.6 ± 11.6 6.9 ± 8.1 20.7 ± 22.4 13.7 ± 22.9 0.060 0.841
MCP-1 (pg/ml) 380 ± 134 288 ± 120 323 ± 102 307 ± 123 0.002 0.023
I-CAM (ng/ml) 360 ± 84 299 ± 64 338 ± 69 328 ± 57 0.001 0.008
V-CAM (ng/ml) 549 ± 101 512 ± 113 567
± 139 536 ± 135 0.060 0.874
E-selectin (ng/ml) 18.9 ± 9.2 12.4 ± 5.0 16.7 ± 5.3 14.4 ± 4.1 0.000 0.014
P-selectin (ng/ml) 125 ± 27 107 ± 27 122 ± 28 112 ± 28 0.002 0.301
L-selectin (ng/ml) 1148 ± 210 1091 ± 197 1081 ± 167 1098 ± 126 0.398 0.128
PAI-1 (ng/ml) 45.0 ± 20.7 29.5 ± 10.1 38.3 ± 11.5 35.2 ± 12.6 0.001 0.026
Values are mean ± SD
P
R
C
F
GE
l
e
s
-E
M
A
C-I
6
-LI
8-LI
les-L
1-PCM
1-IAP
les-P
MA
C-V
FGE
V
-FNT α
-NFI γ
6
.
0
4
.02.0
0
2.
0-
4.
0-
6
.
0-
6-n4:02LPegnahC
tn
e
iciffeo
C
n
o
ital
e
rroC
A
*
*
*
*
*
PRC
FGE
les-E
M
A
C-
I
6
-LI
8-LI
les-L
1-PCM
1-IAP
les-P
MAC-V
FGEV
-FNT
α
-
N
FI
γ
6.
04.
02
.0
02.0
-
4.0
-
6.0
-
3-n5:02/6-n4:02LPegnahC
tneiciffeoCnoitalerroC
B
*
*
*
Fig. 2 Associations between changes in markers of inflammation and changes in PL arachidonic acid (a), and the 20:4n-6/20/5n-3 ratio (b).
Individual bars represent Pearson correlation coefficients *(P \ 0.05)
74 Lipids (2008) 43:65–77
123

Despite the two diet groups consuming roughly the same
caloric intake and all losing at least some weight, there
were larger reductions in the VLCKD group in TNF-a,
IL-8, MCP-1, PAI-1, E-selectin and I-CAM, while these
markers showed little change on low fat suggesting that it
is the macronutrient composition not weight loss or caloric
reduction that is key. Most of the inflammatory markers did
not correlate with weight loss. A correlation would not
have proved that weight loss caused change in inflamma-
tory markers but the lack of correlation makes it extremely
unlikely. As shown in Fig. 3, there is essentially no cor-
relation and if anything the associations tend to go in the
opposite direction of what is expected if weight loss caused
change in markers. In both cases, individuals with the
largest reductions in inflammatory markers tended to be in
the middle of the weight-loss range. The question of weight
loss as a stimulus versus a response has been raised before
with regard to other effects of carbohydrate restriction. Our
group [22] and others [16, 18, 23] have consistently shown
that there is a benefit to atherogenic dyslipidemia, glycemic
control, and insulin from reduction in carbohydrate inde-
pendent of weight loss. We therefore suggest that reduction
in carbohydrate is primary, and weight loss (more precisely
caloric restriction) is not the controlling variable.
One of the most striking responses in fatty acid com-
position was the increase in arachidonic acid and total n-6
PUFA in subjects consuming a VLCKD. Rather than being
a negative factor within lipid membranes, increased ara-
chidonic acid appears to be a beneficial outcome of weight-
reducing diets associated with greater lipolysis [24]. The
increase in plasma arachidonic acid only in response to the
low carbohydrate diet is best explained by decreased de-
gradation presumably due to less interaction with reactive
oxygen species [25]. Increased production from 18:2n-6
was unlikely since the metabolic intermediates 18:3n-6 and
20:3n-6 were reduced in all three circulating fractions and
there was no increase in 20:3n-9, which typically occurs in
cases where PUFA anabolism is increased [13, 25]. Since
arachidonic acid was elevated in all circulating fractions, a
shift from other pools is unlikely. This is supported by the
fact that fat loss was only moderately greater on the
VLCKD compared to the LFD (-5.6 vs. -3.7 kg), and by
week 12, the rate of weight loss on both diets was low.
Given that daily dietary contribution of arachidonic acid is
on the order of 0.5% of the total body pool [26] also
suggests reduced degradation as the major explanation.
Thus, an increase in the proportion of arachidonic acid
resulting from a diet that restricts carbohydrate may be due
to lower catabolism (i.e. better preservation) and therefore
reduced formation of proinflammatory products. The con-
sistent inverse associations between changes in arachidonic
acid and responses in inflammatory markers indicate that
the adverse effects of arachidonic acid are due to meta-
bolites produced subsequent to its release from membranes
rather than the proportion of the intact fatty acid.
The 19% rise in PL arachidonic acid in the low carbo-
hydrate group (from 11.54 to 13.70 wt%) is consistent with
the change seen after 12 weeks of a very low calorie diet in
a more obese population (e.g,, from 9.16 to 11.77 wt%)
[24]. Given the regulatory role of arachidonic acid as a
ligand for PPAR and in gene expression (e.g., fatty acid
synthase) [27], this degree of rise in 20:4n-6 has the
potential to influence fuel partitioning. In the obese Zucker
rat, the increase in liver PL arachidonic acid from 22.68 to
25.23 wt% (an 11% change) induced by feeding 18:3n-6
was associated with significant reductions in food intake
and body fat content [28].
Scenarios associated with less oxidative stress should
result in better preservation of the substrate arachidonic
acid, due to the interaction of free radicals with several
steps in its metabolism. Inflammatory cytokines are known
to increase production of hydroxyl radicals which in turn
initiate arachidonic acid release and breakdown. The
VLCKD in this study resulted in significantly greater
reductions in several proinflammatory markers including
TNF-a, E-selectin, ICAM-1, and IL-8, that were related to
Fig. 3 Associations between
weight loss and changes in
TNF-a and IL-8
Lipids (2008) 43:65–77 75
123

the increase in arachidonic acid. The significantly greater
reduction in TNF-a in subjects following the VLCKD is of
interest in that it is one of the agents known to activate
NF-jB a major transcription factor regulating cytokines,
chemokines and adhesion molecules (TNF-a, MCP-1, IL-8,
E-selectin, and ICAM-1) [12, 29]. The reduction in all of
these agents by the VLCKD suggests that the antiinflam-
matory effects of carbohydrate restriction may be mediated
via down regulation of NF-jB expression [30]. We have
previously found that guinea pigs fed high-cholesterol
atherogenic diets demonstrated significant increases in
aortic TNF-a, an effect that was attenuated by reduction in
dietary carbohydrate [31].
Hypercaloric high carbohydrate feeding stimulates the
production of several fatty acids including 16:0, the major
lipogenic product, and palmitoleic acid (16:1n-7), the
product of D9 desaturase. Palmitoleic acid is a minor
constituent in dietary fat and its increase is a marker of
lipogenesis [32] and its presence has been linked to higher
levels of adiposity [33, 34]. In this study, the VLCKD
resulted in concurrent reductions in both 16:0 and 16:1n-7
in both TG and CE lipid fractions despite an increase in
dietary saturated fat load. The significant reduction in
dietary saturated fat in the LFD led to little decrease in total
saturates and essentially no change in 16:1n-7, with one
subject actually showing a drastic increase. The greater
decrease in circulating SFA in response to carbohydrate
restriction may have contributed to the larger decline in
several inflammatory markers that are regulated by NF-jB
[10, 11]. The decrease in circulating saturated fatty acids
on the VLCKD is likely due to greater oxidation of the
saturated fat from both diet and endogenous lipolysis, and a
reduction in de novo lipogenesis.
In summary, carbohydrate restricted diets that are pro-
portionately high in saturated fatty acids show very
different results from what might be expected [7, 18]. A
VLCKD significantly increases arachidonic acid levels,
presumably due to a better preservation as a result of
reduced oxidative stress and decreased inflammation.
Sparing of arachidonic acid (by reducing its degradation to
oxy-lipids) may provide a signaling mechanism by which
dietary carbohydrate restriction favorably alters lipid
metabolism and inflammatory processes [27].
Acknowledgments This work was supported by the Dr. Robert C.
Atkins Foundation.
References
1. Warensjo E, Riserus U, Vessby B (2005) Fatty acid composition
of serum lipids predicts the development of the metabolic syn-
drome in men. Diabetologia 48:1999–2005
2. Dougherty RM, Galli C, Ferro-Luzzi A, Iacono JM (1987) Lipid
and phospholipid fatty acid composition of plasma, red blood
cells, and platelets and how they are affected by dietary lipids: a
study of normal subjects from Italy, Finland, and the USA. Am J
Clin Nutr 45:443–455
3. King IB, Lemaitre RN, Kestin M (2006) Effect of a low-fat diet
on fatty acid composition in red cells, plasma phospholipids, and
cholesterol esters: investigation of a biomarker of total fat intake.
Am J Clin Nutr 83:227–236
4. Raatz SK, Bibus D, Thomas W, Kris-Etherton P (2001) Total fat
intake modifies plasma fatty acid composition in humans. J Nutr
131:231–234
5. Phinney SD (2006) The low fat paradox––do dietary carbohy-
drates increase circulating saturated fatty acids? Am J Clin Nutr
84:461–462
6. Parks EJ (2002) Changes in fat synthesis influenced by dietary
macronutrient content. Proc Nutr Soc 61:281–286
7. Volek JS, Feinman RD, Phinney SD, Forsythe CE, Silvestre R,
Judelson D, Quann EE, Wood RJ, Puglisi MJ, LaBonte CC,
Kraemer WJ, Bibus DM, Contois JH, Fernandez ML (2007)
Comparative effects of dietary restriction of carbohydrate or fat
on circulating saturated fatty acids and atherogenic dyslipidemia.
(Submitted)
8. Reaven GM (1988) Banting lecture 1988. Role of insulin resis-
tance in human disease. Diabetes 37:1595–1607
9. Dandona P, Aljada A, Chaudhuri A, Mohanty P, Garg R (2005)
Metabolic syndrome: a comprehensive perspective based on
interactions between obesity, diabetes, and inflammation. Circu-
lation 111:1448–1454
10. Lee JY, Hwang DH (2006) The modulation of inflammatory gene
expression by lipids: mediation through Toll-like receptors. Mol
Cells 21:174–185
11. Lee JY, Zhao L, Youn HS, Weatherill AR, Tapping R, Feng L,
Lee WH, Fitzgerald KA, Hwang DH (2004) Saturated fatty acid
activates but polyunsaturated fatty acid inhibits Toll-like receptor
2 dimerized with Toll-like receptor 6 or 1. J Biol Chem
279:16971–16979
12. de Winther MP, Kanters E, Kraal G, Hofker MH (2005) Nuclear
factor kappaB signaling in atherogenesis. Arterioscler Thromb
Vasc Biol 25:904–914
13. Calder PC (2006) Polyunsaturated fatty acids and inflammation.
Prostaglandins Leukot Essent Fatty Acids 75(3):197–202
14. Johnston CS, Tjonn SL, Swan PD, White A, Hutchins H, Sears B
(2006) Ketogenic low-carbohydrate diets have no metabolic
advantage over nonketogenic low-carbohydrate diets. Am J Clin
Nutr 83:1055–1061
15. Peplow PV (1996) Actions of cytokines in relation to arachidonic
acid metabolism and eicosanoid production. Prostaglandins
Leukot Essent Fatty Acids 54:303–317
16. Feinman RD, Volek JS (2006) Low carbohydrate diets improve
atherogenic dyslipidemia even in the absence of weight loss. Nutr
Metab (Lond) 3:24
17. Feinman RD, Westman EC, Volek JS (2006) Low carbohydrate
and low fat diets in diabetes, cardiovascular disease and meta-
bolic syndrome. Cellscience Reviews 3, no. 1 (http://www.
cellscience.com/reviews9/Low_carbohydrate_low_fat_diets_in_
disease. html)
18. Krauss RM, Blanche PJ, Rawlings RS, Fernstrom HS, Williams
PT (2006) Separate effects of reduced carbohydrate intake and
weight loss on atherogenic dyslipidemia. Am J Clin Nutr
83:1025–1031
19. Volek JS, Feinman RD (2005) Carbohydrate restriction improves
the features of metabolic syndrome. Metabolic syndrome may be
defined by the response to carbohydrate restriction. Nutr Metab
(Lond) 2:31
20. Kasim-Karakas SE, Tsodikov A, Singh U, Jialal I (2006)
Responses of inflammatory markers to a low-fat, high-carbohy-
drate diet: effects of energy intake. Am J Clin Nutr 83:774–779
76 Lipids (2008) 43:65–77
123

21. Liu S, Manson JE, Buring JE, Stampfer MJ, Willett WC, Ridker
PM (2002) Relation between a diet with a high glycemic load and
plasma concentrations of high-sensitivity C-reactive protein in
middle-aged women. Am J Clin Nutr 75:492–498
22. Volek JS, Sharman MJ, Gomez AL, Scheett TP, Kraemer WJ
(2003) An isoenergetic very low carbohydrate diet improves
serum HDL cholesterol and triacylglycerol concentrations, the
total cholesterol to HDL cholesterol ratio and postprandial
pipemic responses compared with a low fat diet in normal weight,
normolipidemic women. J Nutr 133:2756–2761
23. Gannon MC, Nuttall FQ (2006) Control of blood glucose in type
2 diabetes without weight loss by modification of diet composi-
tion. Nutr Metab (Lond) 3:16
24. Phinney SD, Davis PG, Johnson SB, Holman RT (1991) Obesity
and weight loss alter serum polyunsaturated lipids in humans. Am
J Clin Nutr 53:831–838
25. Sullivan PG, Rippy NA, Dorenbos K, Concepcion RC, Agarwal
AK, Rho JM (2004) The ketogenic diet increases mitochondrial
uncoupling protein levels and activity. Ann Neurol 55:576–580
26. Zhou L, Nilsson A (2001) Sources of eicosanoid precursor fatty
acid pools in tissues. J Lipid Res 42:1521–1542
27. Brash AR (2001) Arachidonic acid as a bioactive molecule. J Clin
Invest 107:1339–1345
28. Phinney SD, Tang AB, Thurmond DC, Nakamura MT, Stern JS
(1993) Abnormal polyunsaturated lipid metabolism in the obese
Zucker rat, with partial metabolic correction by gamma-linolenic
acid administration. Metabolism 42:1127–1140
29. Monaco C, Paleolog E (2004) Nuclear factor kappaB: a potential
therapeutic target in atherosclerosis and thrombosis. Cardiovasc
Res 61:671–682
30. Shin WS, Szuba A, Rockson SG (2002) The role of chemokines
in human cardiovascular pathology: enhanced biological insights.
Atherosclerosis 160:91–102
31. Fernandez ML, Volek JS (2006) Guinea pigs: a suitable animal
model to study lipoprotein metabolism, atherosclerosis and
inflammation. Nutr Metab (Lond) 3:17
32. Aarsland A, Wolfe RR (1998) Hepatic secretion of VLDL fatty
acids during stimulated lipogenesis in men. J Lipid Res 39:1280–
1286
33. Okada T, Furuhashi N, Kuromori Y, Miyashita M, Iwata F, Ha-
rada K (2005) Plasma palmitoleic acid content and obesity in
children. Am J Clin Nutr 82:747–750
34. Kunesova M, Hainer V, Tvrzicka E, Phinney SD, Stich V,
Parizkova J, Zak A, Stunkard AJ (2002) Assessment of dietary
and genetic factors influencing serum and adipose fatty acid
composition in obese female identical twins. Lipids 37:27–32
Lipids (2008) 43:65–77 77
123