Content uploaded by Cristina Andres-Lacueva
Author content
All content in this area was uploaded by Cristina Andres-Lacueva
Content may be subject to copyright.
Relationship of Plasma Polyunsaturated Fatty Acids to
Circulating Inflammatory Markers
Luigi Ferrucci, Antonio Cherubini, Stefania Bandinelli, Benedetta Bartali, Annamaria Corsi,
Fulvio Lauretani, Antonio Martin, Cristina Andres-Lacueva, Umberto Senin, and Jack M. Guralnik
Longitudinal Studies Section (L.F.), Clinical Research Branch, National Institute on Aging, Baltimore, Maryland 21225;
Department of Clinical and Experimental Medicine (A.Ch., U.S.), Institute of Gerontology and Geriatrics, Perugia University
Medical School, 06100 Perugia, Italy; Department of Geriatric Rehabilitation (S.B.), Tuscany Regional Health Agency
(A.Co., F.L.), 50125 Florence, Italy; Division of Nutritional Sciences (B.B.), Cornell University, Ithaca, New York 14853;
Human Nutrition Research Center on Aging (A.M.), Tufts University, Boston, Massachusetts 02111; Department of Nutrition
and Food Science (C.A.-L.), CeRTA, University of Barcelona, 08028 Barcelona, Spain; and Laboratory of Epidemiology,
Demography, and Biometry (J.M.G.), National Institute on Aging, Bethesda, Maryland 20892
Aims: Persons with high intake of polyunsaturated fatty acids
(PUFAs) have lower cardiovascular morbidity and mortality. The
protective effect of PUFAs is mediated by multiple mechanisms, in-
cluding their antiinflammatory properties. The association of phys-
iological PUFA levels with pro- and antiinflammatory markers has
not been established.
Methods and Results: In 1123 persons (aged 20 –98 yr), we exam-
ined the relationship between relative concentration of fatty acids in
fasting plasma and level of inflammatory markers. Adjusting for age,
sex, and major confounders, lower arachidonic and docosahexaenoic
acids were associated with significantly higher IL-6 and IL-1ra and
significantly lower TGF

. Lower
␣
-linolenic acid was associated with
higher C-reactive protein and IL-1ra, and lower eicosapentaenoic acid
was associated with higher IL-6 and lower TGF

. Lower docosa-
hexaenoic acid was strongly associated with lower IL-10. Total n-3
fatty acids were associated with lower IL-6 (P ⫽ 0.005), IL-1ra (P ⫽
0.004), and TNF
␣
(P ⫽ 0.040) and higher soluble IL-6r (P ⬍ 0.001),
IL-10 (P ⫽ 0.024), and TGF

(P ⫽ 0.0012). Lower n-6 fatty acid levels
were significantly associated with higher IL-1ra (P ⫽ 0.026) and lower
TGF

(P ⫽ 0.014). The n-6 to n-3 ratio was a strong, negative correlate
of IL-10. Findings were similar in participants free of cardiovascular
diseases and after excluding lipids from covariates.
Conclusions: In this community-based sample, PUFAs, and espe-
cially total n-3 fatty acids, were independently associated with lower
levels of proinflammatory markers (IL-6, IL-1ra, TNF
␣
, C-reactive
protein) and higher levels of antiinflammatory markers (soluble IL-
6r, IL-10, TGF

) independent of confounders. Our findings support
the notion that n-3 fatty acids may be beneficial in patients affected
by diseases characterized by active inflammation. (J Clin Endocri-
nol Metab 91: 439– 446, 2006)
T
HERE IS EVIDENCE that a diet rich in polyunsaturated
fatty acids (PUFAs) and, in particular, the omega-3
family (n-3), is associated with lower cardiovascular mor-
bidity and mortality and reduced risk of sudden death, in-
dependent of other known cardiovascular risk factors (1–5).
Studies have suggested that the protective effects of n-3
PUFA are mediated by multiple mechanisms, including their
antiinflammatory properties (6).
Preclinical studies have shown that fatty acids modulate
the inflammatory response by multiple mechanisms, includ-
ing transcriptional down-regulation of proinflammatory cy-
tokines and the vascular surface expression of endothelial
leukocyte adhesion molecules (7, 8). In particular, in exper-
imental and animal models, n-3 fatty acids inhibit the pro-
duction of IL-1 and TNF
␣
(7). An antiinflammatory effect of
n-3 fatty acids is supported by the beneficial effect of n-3
fatty-acid supplementation in patients affected by diseases
characterized by active inflammation, such as rheumatoid
arthritis and Crohn’s disease (9). In small groups of healthy
volunteers, dietary supplementation with n-3 fatty acids was
associated with reduced levels of IL-1

, thromboxane

2
, and
prostaglandin E2 (10, 11) but not C-reactive protein (CRP) (12).
A recent study showed that dietary intake of n-3 and
omega-6 (n-6) fatty acids in American men and women was
inversely associated with plasma levels of soluble TNF
␣
receptors 1 and 2 but not with other cytokines (13). This
observation is important because it suggests that physiolog-
ical levels of fatty acids modulate inflammation. However,
studies on dietary intake should be complemented by studies
that investigate the relationship between fatty acid plasma
levels and serum levels of multiple inflammatory markers. In
fact, although some studies found a good correlation be-
tween dietary intake of fatty acids and blood levels (14),
others did not confirm this finding, reporting only a modest
association (15), likely due to the fact that circulating fatty
acid levels reflect the interplay among dietary intake, ab-
sorption, and metabolism.
Using data from a representative sample of the general
population, we tested the hypothesis that circulating levels
First Published Online October 18, 2005
Abbreviations: AA, Arachidonic acid; ALA,
␣
-linolenic acid; BMI,
body mass index; CRP, C-reactive protein; DHA, docosahexaenoic acid;
EPA, eicosapentaenoic acid; FAME, fatty acid methyl esters; HDL, high-
density lipoprotein; IL-1ra, IL-1 receptor antagonist; InCHIANTI, In-
vecchiare in Chianti, aging in the Chianti area; LA, linoleic acid; LDL,
low-density lipoprotein; n-3, omega-3 family; n-6, omega-6 family;
PPaR, peroxisome proliferator activated receptor; PUFA, polyunsatu-
rated fatty acid; sIL-6r, soluble IL-6 receptor.
JCEM is published monthly by The Endocrine Society (http://www.
endo-society.org), the foremost professional society serving the en-
docrine community.
0021-972X/06/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 91(2):439 – 446
Printed in U.S.A. Copyright © 2006 by The Endocrine Society
doi: 10.1210/jc.2005-1303
439
on February 24, 2006 jcem.endojournals.orgDownloaded from
of selected PUFAs are associated with lower concentrations
of proinflammatory cytokines and, possibly, higher levels of
antiinflammatory cytokines. This information is important in
furthering our understanding of the mechanisms by which
fatty acids modulate cardiovascular risk and other clinical
conditions characterized by a proinflammatory state.
Subjects and Methods
Study population and collection of blood samples.
Invecchiare in Chianti, aging in the Chianti area (InCHIANTI) is an
epidemiological study conducted in two small towns of Tuscany, Italy.
The rationale, design, and data collection methods of InCHIANTI are
described elsewhere (16). In brief, in August 1998, 1270 persons aged 65
yr or older were randomly selected from the population registry of the
two sites. Additionally, men and women randomly sampled from the
age strata 20 –29, 30 –39, 40 –49, 50–59, and 60– 64 yr were sequentially
invited to participate in the study, until at least 30 men and 30 women
in each decade, ages 20– 69 yr, had been enrolled.
Of the 1714 eligible persons, 640 men and 813 women (84.8%) agreed
to participate and were interviewed. Of those, 595 men and 748 women
(92.4%) provided blood samples. Data on plasma fatty acid and serum
cytokine composition were obtained for 1180 participants (87.9%). Be-
cause poor cognitive status is associated with inflammatory conditions
and strongly affects dietary intake (17), we also excluded 57 participants
in whom we established a dementia diagnosis, based on the Diagnostic
and Statistical Manual of Mental Disorders, version III-R criteria (18).
Thus, the final study population included 1123 participants, none of
whom had dietary supplementation of fatty acids.
The study protocol complies with the Declaration of Helsinki and was
approved by the Italian National Institute of Research and Care on
Aging Ethical Committee. Participants received an extensive description
of the study and signed an informed participation consent that included
permission to conduct analyses on the biological specimens collected
and stored.
Laboratory analysis
Blood samples were collected in the morning after a 12-h fast. Ali-
quots of serum and plasma were immediately obtained and stored at
⫺80 C. The samples used to measure circulating levels of cytokines and
fatty acids had not been previously thawed.
Fatty acids were measured using a fasting plasma sample. A known
amount of heptadecanoic acid (17:0) (Sigma Chemical Co., St Louis, MO)
was added to each sample as an internal standard, and total lipids were
extracted from 0.15 ml of plasma (19). In a pilot study, we had found that
no traces of heptadecanoic acid were detectable in 25 plasma samples
from InCHIANTI participants. Fatty acid methyl esters (FAME) were
prepared through transesterification using Lepage and Roy’s method
(20), modified according to Rodriguez-Palmero et al. (21). Separation of
FAME was carried out on an HP-6890 gas chromatograph (Hewlett-
Packard, Palo Alto, CA) with a 30-m fused silica column (HP-225;
Hewlett-Packard). FAMEs were identified by comparison with pure
standards (NU Chek Prep, Inc., Elysian, MA). For quantitative analysis
of fatty acids as methyl esters, calibration curves for FAME (ranging
from C14:0 to C24:1) were prepared by adding six increasing amounts
of individual FAME standards to the same amount of internal standard
(C17:0; 50
g). The correlation coefficients for the calibration curves of
20 fatty acids were in all cases higher than 0.998 in the range of con-
centrations studied. Fatty acid concentration was expressed as a per-
centage. Fatty acid percent area/area was also calculated. The coefficient
of variation for all fatty acids was on average 1.6% for intraassay and
3.3% for interassay.
In the present analysis, we examined data on the concentration of
PUFAs, which are characterized by two or more double bonds in the
hydrocarbon chain. The n-3 and n-6 families of fatty acids account for
more than 95% of total PUFAs and are named from the position of the
first double bond, located on the third or sixth carbon, respectively, from
the terminal methyl group (22). The total n-6 fatty acids included linoleic
(LA) (C18:2-n6), eicosadienoic (C20:2n6), dihomo-g-linolenic (C20:3n6),
and arachidonic (AA) (C20:4n6) acids, whereas the total n-3 fatty acids
included
␣
-linolenic (ALA) (C18:3n3), eicosapentaenoic (EPA) (C20:
5n3), and docosahexaenoic (C22:6-n3) (DHA) acids.
Serum levels of IL-6, soluble IL-6 receptor (sIL-6r, 80 kDa), IL-1

, IL-1
receptor antagonist (IL-1ra), and TNF
␣
were measured by ELISAs using
commercial kits (BIOSOURCE International, Camarillo, CA). TGF

and
IL-10 levels were measured in duplicate using highly sensitive quanti-
tative sandwich assays (Quantikine HS, R&D Systems, Minneapolis,
MN). The lowest detectable concentrations were 0.1 pg/ml for IL-6, 8
pg/ml for sIL-6r, 0.01 pg/ml for IL-1

, 0.09 pg/ml for TNF
␣
, 4 pg/ml
for IL-1ra, 7 pg/ml for TGF

, and 1.5 pg/ml for IL-10. The interassay
coefficient of variation was 4.5% for IL-1ra and less than 8.0% for the
other cytokines.
CRP was measured in duplicate using an ELISA colorimetric com-
petitive immunoassay that used purified protein and polyclonal anti-
CRP antibodies. The minimum detectable concentration was 0.03 mg/
liter and the interassay coefficient of variation was 5.0%. Total
cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycer-
ides were assessed using commercial enzymatic tests (Roche Diagnos-
tics, Mannheim, Germany).
Covariates
Participants were classified as nonsmokers or former smokers vs.
current smokers based on self-report. Weight was measured using a
high-precision mechanical scale. Standing height was measured to the
nearest 0.1 cm. Body mass index (BMI) was calculated as weight (kilo-
grams)/height (square meters). Average daily intake of energy (kilo-
calories) and carbohydrates, proteins, total lipids and unsaturated, and
monosaturated and polyunsaturated fatty acids (grams) were estimated
by administering the European Prospective Investigation into Cancer
and Nutrition food frequency questionnaire, which has been extensively
validated in the Italian population (23).
Study participants responded to an extensive questionnaire on ha-
bitual physical activity and were classified as sedentary if they reported
being completely inactive or performing low-intensity physical activity,
such as short walking or light housekeeping activities totaling less than
2 h/wk.
A physician evaluated all participants. Diseases were ascertained
according to standard, preestablished criteria that combined informa-
tion from self-reported physician diagnoses, current pharmacological
treatment, medical records, clinical examinations, and blood tests. Dis-
eases included in the current analysis were coronary heart disease (in-
cluding angina and myocardial infarction), congestive heart failure,
cerebrovascular disease (including transient ischemic attack and stroke),
diabetes, and hypertension. Diagnostic algorithms were modified ver-
sions of those created for the Women’s Health and Aging Study (24). An
ankle-arm index of 0.9 or less was considered indicative of peripheral
arterial disease (25). Participants were asked to report all drugs taken at
least once over the last 15 d. Using this information, we created a
dichotomous variable indicating whether the participant received treat-
ments that may affect circulating levels of fatty acids and/or inflam-
matory markers, including statins, other drugs aimed at reducing cir-
culating lipids, steroids, nonsteroidal antiinflammatory drugs, and
angiotensin-converting enzyme inhibitors. This condition is defined in
the text as potentially confounding drug treatment.
Statistical analysis
Continuous variables are reported as mean ⫾ sd and categorical
variables as percentages. Log-transformed values of cytokines, except
for TGF

, were used in the analysis.
The relationships of specific fatty acids with potential covariates were
explored by computing age- and sex-adjusted partial Pearson correla-
tions. Further analyses were performed to test the mutually independent
effects of total n-3 and n-6 fatty acids on inflammatory markers, after
adjusting for age; sex; education; intake of energy, proteins, and car-
bohydrates; physical activity; BMI; smoking; low-density lipoprotein
(LDL) cholesterol; HDL cholesterol; triglycerides; hypertension; diabe-
tes; coronary heart disease; congestive heart failure; stroke; peripheral
arterial disease; and potentially confounding drug treatment. All anal-
yses were performed using the SAS statistical package (version 9.1; SAS
Institute, Inc., Cary, NC) with a statistical significance level set at P ⬍
0.05.
440 J Clin Endocrinol Metab, February 2006, 91(2):439–446 Ferrucci et al. • PUFA Levels Correlate with Low Inflammation
on February 24, 2006 jcem.endojournals.orgDownloaded from
Results
The principal characteristics of the study population are
reported in the first column of Table 1. Only data on total n-3
and n-6 PUFAs are shown. Both total n-3 and n-6 were
negatively correlated with age (n-3: r ⫽⫺0.13; n-6: r ⫽⫺0.33)
but not sex (n-3: r ⫽ 0.02; n-6: r ⫽ 0.03). Independent of age
and sex, total n-3 PUFAs were positively correlated with
education and HDL cholesterol and negatively correlated
with triglycerides. Total n-6 PUFAs were positively corre-
lated with total, LDL, and HDL cholesterol and negatively
correlated with most of the other cardiovascular risk factors
and cardiovascular diseases (Table 1). LA was positively and
independently correlated with LDL cholesterol. LA (r ⫽
0.31), AA (r ⫽ 0.28), EPA (r ⫽ 0.23), and DHA (r ⫽ 0.14) were
positively correlated with HDL cholesterol. LA (r ⫽ 0.51) was
positively correlated (r ⫽ 0.51), whereas AA (r ⫽⫺0.34), EPA
(r ⫽⫺0.21), and DHA (r ⫽⫺0.20) were negatively correlated
with triglycerides. Other correlations of specific fatty acids
with the variables reported in Table 1 were small (⬍0.09) and
generally not statistically significant. In particular, neither
n-3 nor n-6 fatty acids were independently correlated with
parameters of dietary intake.
The mean plasma concentration of total fatty acids was
3200 ⫾ 724 mg/liter (range 1295– 6885), which is compatible
with those reported in a group of middle-aged American
women and blood donors (26, 27). The concentration tended to
be higher at older ages (r ⫽ 0.08, P ⬍ 0.0023) with no substantial
difference between men and women (men, 3155 ⫾ 777 vs.
women, 3236 ⫾ 677 mg/liter, P ⫽ 0.07). As expected, the plasma
concentration of total fatty acids was positively correlated with
total cholesterol (r ⫽ 0.59, P ⬍ 0.0001) and triglycerides (r ⫽ 0.71,
P ⬍ 0.0001) and negatively correlated with HDL cholesterol (r ⫽
⫺0.07, P ⬍ 0.025). Total fatty acids were not independently
correlated with any of the inflammatory markers considered in
this study. n-3 and n-6 fatty acids accounted for 3.4 and 33.1%
of total fatty acids. The percentage of n-3 and n-6 fatty acids on
total fatty acids was significantly lower in older participants,
respectively, by 0.002% per year (P ⬍ 0.0001) for n-3 and 0.008%
per year (P ⬍ 0.0001) for n-6, without substantial difference
between men and women.
Of the plasma n-3 fatty acids, 13.6 ⫾ 6.1% were ALA,
18.8 ⫾ 5.0% were EPA and 67.7 ⫾ 4.9% were DHA. Of the
n-6 fatty acids, 75.3 ⫾ 5.0% were LA and 24.4 ⫾ 4.9% were
AA. Levels of inflammatory markers according to quartiles
of specific fatty acids are reported in Table 2. All mean values
and statistical tests are adjusted for age and multiple con-
founders. Lower AA and DHA were associated with higher
IL-6 and IL-1ra and lower TGF

. Lower DHA was also as-
sociated with lower IL-10. Lower ALA was associated with
higher CRP and IL-1ra, and lower EPA was associated with
higher IL-6, lower TGF

, and lower IL-10. Participants in the
two lower quartiles of LA had significantly lower sIL-6r than
those in the two upper quartiles. AA/EPA ratio was not
associated with any of the different inflammatory markers.
After removing lipids as covariates from these models, these
results were substantially unchanged, except that the inverse
association of IL-1ra with LA (adjusted values across quar-
tiles: 149, 129,and 127 pg/ml; P ⫽ 0.0003) and EPA (143, 132,
128, and 130; P ⫽ 0.0076) became stronger and highly statisti-
cally significant (cf. Table 2). Restricting the analysis to the 432
TABLE 1. Characteristics of the study population (n ⫽ 1123) and their correlation with n-3 and n-6 fatty acids
Mean ⫾ SD or n (%)
Age and sex adjusted partial correlations r (P value)
With total n-3 FA
a
With total n-6 FA
Age (yr) 68.2 ⫾ 15.4
⬍65 246 (21.9)
65–74 504 (44.9)
75–84 276 (24.6)
ⱖ85 97 (8.6)
Sex (women) 620 (55.2)
Years in school 6.6 ⫾ 4.2 0.14 (⬍0.0001) 0.05 (0.13)
BMI (kg/m
2
) 27.5 ⫾ 4.4 ⫺0.04 (0.13) ⫺0.14 (⬍0.0001)
Energy intake (kcal/d) 2027 ⫾ 621 ⫺0.07 (0.03) ⫺0.02 (0.47)
Carbohydrate intake (g/d) 261 ⫾ 87 ⫺0.06 (0.03) ⫺0.01 (0.69)
Protein intake (g/d) 79 ⫾ 23 ⫺0.05 (0.10) ⫺0.02 (0.46)
Total lipids intake (g/d) 69 ⫾ 23 ⫺0.05 (0.08) ⫺0.01 (0.86)
Saturated FA intake (g/d) 23 ⫾ 9 ⫺0.05 (0.09) 0.02 (0.45)
Monounsaturated FA intake (g/d) 35 ⫾ 12 ⫺0.02 (0.46) ⫺0.03 (0.29)
Polyunsaturated FA intake (g/d) 7 ⫾ 2 ⫺0.05 (0.11) 0.01 (0.82)
Total cholesterol (mg/dl) 215 ⫾ 40 0.00 (0.95) 0.11 (⬍0.001)
LDL cholesterol (mg/dl) 134 ⫾ 35 0.02 (0.58) 0.18 (⬍0.0001)
HDL cholesterol (mg/dl) 56 ⫾ 15 0.14 (⬍0.0001) 0.38 (⬍0.0001)
Triglycerides (mg/dl) 123 ⫾ 65 ⫺0.21 (⬍0.0001) ⫺0.58 (⬍0.0001)
Current smoker 179 (16) ⫺0.03 (0.36) ⫺0.07 (0.030)
Sedentary 186 (17) ⫺0.06 (0.06) ⫺0.09 (0.004)
Coronary heart disease 58 (5.2) ⫺0.01 (0.69) ⫺0.06 (0.020)
Stroke 21 (1.9) ⫺0.02 (0.46) ⫺0.08 (0.003)
Congestive heart failure 62 (5.5) ⫺0.07 (0.02) ⫺0.09 (0.004)
Hypertension 500 (45.5) ⫺0.04 (0.21) ⫺0.08 (0.005)
Diabetes 78 (7.0) 0.01 (0.81) 0.03 (0.24)
Peripheral artery disease 139 (12.6) ⫺0.04 (0.15) ⫺0.08 (0.010)
FA, Fatty acids.
a
Log-transformed values were used for the correlation analysis.
Ferrucci et al. • PUFA Levels Correlate with Low Inflammation J Clin Endocrinol Metab, February 2006, 91(2):439–446 441
on February 24, 2006 jcem.endojournals.orgDownloaded from
TABLE 2. Multivariate analysis of the relationship between specific n-3 and n-6 fatty acids and inflammatory markers
LA (quartiles), %
a
AA (quartiles), %
a
ALA (quartiles), %
a
Quartiles limits ⬍22.33 22.33–24.95 24.96–27.51 ⬎27.51
ref
P for
trend
⬍6.82 6.82–7.97 7.98 –9.28 ⬎9.28
ref
P for
trend
⬍0.304 0.304–0.381 0.382– 0.493 ⬎0.493
ref
P for
trend
Median 20.5 23.5 26.2 29.4 5.95 7.34 8.58 10.19 0.264 0.344 0.427 0.640
IL-6 (pg/ml) 1.32 1.23 1.28 1.28 0.38 1.44
b
1.31
b
1.28 1.15 0.0024 1.35 1.28 1.31 1.20 0.12
sIL-6r (ng/ml) 83
c
86
c
90 97 0.06 84
b
80 87 96 0.1452 87 87 90 92 0.50
IL-1

(pg/ml) 0.13 0.14 0.13 0.14 0.60 0.13 0.14 0.14 0.13 0.52 0.14 0.12 0.14 0.13 0.41
IL-1ra (pg/ml) 140 127 129 131 0.26 141
c
140
c
126 126 0.0008 143
c
127 132 128 0.0248
TNF-
␣
(pg/ml) 4.8 4.7 4.7 4.8 0.67 4.8 5.0 4.5 4.7 0.29 5.1 4.4 4.7 4.8 0.53
IL-10 (pg/ml) 3.3 3.6 4.8 3.5 0.91 3.4 3.7 4.4 3.5 0.98 3.3 3.7 3.8 3.9 0.17
TGF-

(ng/ml) 11.20
c
11.90 12.00 12.60 0.38 10.60
b
11.40 12.80 12.90 0.0003 11.8 11.7 11.5 12.7 0.35
CRP (mg/liter) 2.70 2.50 2.50 2.40 0.90 2.50 2.70
c
2.70
c
2.30 0.60 3.0
b
2.6 2.3 2.4 0.0047
EPA (quartiles), %
a
DHA (quartiles), %
a
AA/EPA ratio (quartiles)
Quartiles limits ⬍0.491 0.491–0.585 0.586 –0.707 ⬎0.707
ref
P for
trend
⬍1.77 1.77–2.23 2.24–2.75 ⬎2.75
ref
P for
trend
⬍11.40 11.40–13.20 13.21–15.95 ⬎15.95
ref
P for
trend
Median 0.422 0.543 0.647 0.800 1.46 2.00 2.47 3.18 9.82 12.28 14.50 18.16
IL-6 (pg/ml) 1.37
b
1.41
b
1.24 1.16 0.0033 1.43
b
1.30 1.25 1.18 0.0075 1.22 1.34 1.31 1.25 0.51
sIL-6r (ng/ml) 86
c
88 89 94 0.39 85
c
87 91 94 0.19 89 88 84
c
95 0.75
IL-1

(pg/ml) 0.13 0.14 0.14 0.13 0.85 0.13 0.14 0.13 0.13 0.61 0.13 0.14 0.14 0.14 0.44
IL-1ra (pg/ml) 138 132 129 131 0.20 143
b
130 135 125 0.0021 131 136 133 133 0.59
TNF-
␣
(pg/ml) 5.2 4.6 4.5 4.7 0.13 4.8 5.1 4.7 4.5 0.06 4.9 4.4 4.8 4.8 0.69
IL-10 (pg/ml) 3.0 3.7 4.0 4.1 0.07 2.9 3.8 3.6 4.7 0.003 4.4 3.5 3.9 3.3 0.11
TGF-

(ng/ml) 11.40
b
11.00
b
11.90
c
13.10 0.0071 10.90
b
11.50
c
11.90
c
13.20 0.0004 11.6 12.3 11.5 12.4 0.79
CRP (mg/liter) 2.60 2.60 2.40 2.50 0.97 2.80
c
2.40 2.60 2.30 0.17 2.5 2.6 2.6 2.4 0.55
ref, Reference quartile.
a
Mean values and statistics for the association of specific fatty acids with selected inflammatory markers are adjusted for age, sex, education, daily intake of energy, carbohydrates,
proteins and lipids, physical activity, BMI, smoking, LDL cholesterol, HDL cholesterol, triglycerides, hypertension, diabetes, coronary heart disease, congestive heart failure, stroke,
peripheral artery disease, and potentially confounding drug treatment.
b
P ⬍ 0.01, compared with the reference quartile.
c
P ⬍ 0.05, compared with the reference quartile.
442 J Clin Endocrinol Metab, February 2006, 91(2):439–446 Ferrucci et al. • PUFA Levels Correlate with Low Inflammation
on February 24, 2006 jcem.endojournals.orgDownloaded from
men and 569 women free of prevalent cardiovascular disease,
all the associations that were statistically significant in Table 2
remained statistically significant, and in addition, the inverse
relationship between LA and sIL-6r and between DHA and
TNF
␣
that were borderline statistically significant became sta-
tistically significant (respectively, P ⫽ 0.020 and P ⫽ 0.040).
Figure 1 shows median serum levels of inflammatory mark-
ers according to n-3 and n-6 quartiles. Adjusting for age and sex,
lower total n-3 and n-6 PUFAs were associated with higher IL-6,
IL1-ra, TNF
␣
, and CRP and lower IL-6r, IL-10, and TGF

levels.
In most cases, the associations were highly statistically signif-
icant, with an evident dose-response relationship.
In subsequent models predicting inflammatory markers, we
simultaneously entered both total n-3 and total n-6 PUFAs as
well as multiple potential confounders and tested n-6 to n-3
ratio as predictor. From these models, we estimated the mean
values of inflammatory markers, according to n-3 and n-6 quar-
tiles and n-6 to n-3 ratio quartiles, which are reported and
statistically compared in Table 3. Lower total n-3 PUFAs were
still strongly and significantly associated with higher IL-6, IL-
1ra, and TNF
␣
and lower total sIL-6r, IL-10, and TGF

but no
longer with higher CRP. Lower n-6 PUFAs were independently
associated with higher IL-1ra and lower TGF

but no longer
with the other inflammatory markers. The n-6 to n-3 ratio was
positively associated with IL-6 and IL-1ra and, based on P
value, was the strongest negative correlate of IL-10 and TGF

.
These findings were confirmed in analyses performed sepa-
rately in men and women, in participants 65 yr and older and
after restricting the study population to the 1001 participants
free of coronary artery disease, congestive heart failure, stroke,
and peripheral arterial disease.
Discussion
Our findings are consistent with the hypothesis that n-3
fatty acids have antiinflammatory properties (6, 7, 11, 28).
IL-6 and TNF
␣
are generally considered proinflammatory
cytokines and the potent antiinflammatory properties of
IL-10 and TGF

are well known (29, 30). The negative as-
sociation of total n-3 fatty acids with IL-1ra (a competitive
inhibitor of the proinflammatory cytokine IL-1) and the pos-
itive association with sIL-6r (that in certain conditions en-
hances the biological activity of IL-6) requires discussion. At
a molecular level, IL-1ra is a natural antagonist of the proin-
flammatory cytokine IL-1 (31), and in animal models of
chronic inflammation, the administration of IL-1ra prevents
tissue damage (32). Despite this, as a circulating biomarker,
IL-1ra is considered an acute-phase protein and a more re-
liable measure of proinflammatory state than IL-1 (33). Sim-
ilar triggers induce the production of IL-1 and IL-1ra, but IL-1
is produced locally and only small quantities spill in the serum,
whereas IL-1ra is produced by the liver in large quantities and
fully released into the circulation. For example, in experimental
endotoxemia in humans, IL-1 increases in the circulation only
by a factor of 2–2.5, whereas IL-1ra increases by a factor of 10 –20
(32, 34). This may also explain why we did not find any asso-
ciation between fatty acids and IL-1

in our study.
The role of sIL-6r in inflammation is still unclear. In gen-
FIG. 1. Median serum levels of inflammatory markers according to total n-3 and n-6 PUFAs quartiles. Comparisons between groups and tests
for trend are based on age- and sex-adjusted nonparametric ANOVA. The thresholds for quartile definition were 2.72, 3.27, and 3.93 for total
n-3 fatty acids and 30.4, 33.13, and 36.26 for total n-6 fatty acids.
Ferrucci et al. • PUFA Levels Correlate with Low Inflammation J Clin Endocrinol Metab, February 2006, 91(2):439–446 443
on February 24, 2006 jcem.endojournals.orgDownloaded from
eral, a specific IL-6 receptor is expressed by the membranes
of hepatocytes, monocytes/macrophages, and some leuko-
cytes. However, the sIL-6r/IL-6 complex can stimulate a
much wider range of cell types that have the gp130 protein,
the inner portion of the IL-6 receptor, on their membrane.
Recent literature suggests that this mechanism is active only
when plasma levels of soluble gp130 are low, which is prob-
ably a rare condition. When high levels of gp130 and sIL-6r
are present, a hexameric complex is created (2*IL-6 ⫹ 2*sIL-6r
⫹ 2*gp130), which tends to precipitate and has no biological
activity (35). Thus, in this condition, sIL-6r levels are anti-
inflammatory (36). IL-6 and sIL-6r have shown opposite bi-
ological activity in several instances. Recently we reported
that IL-6 is associated with insulin resistance, whereas sIL-6r
has the opposite effect (37).
Similarly to Pischon et al. (13), we found some evidence that
n-6 fatty acids may be antiinflammatory with no evidence of the
proinflammatory activity previously suggested by many au-
thors. However, our data suggest that the immunomodulatory
effect of PUFAs may be influenced by the n-6 to n-3 ratio, which
in our study was the strongest negative correlate of IL-10 and
TGF

, two powerful antiinflammatory cytokines. Pischon et al.
found no significant association between plasma PUFA con-
centrations and serum CRP levels, and in our study only ALA
was an independent negative correlate of CRP. This particular
finding is somewhat puzzling and difficult to interpret because
the production of CRP is mainly regulated by IL-6 (38), and
ALA was not a significant independent correlate of IL-6. In-
terestingly, in a recent trial, supplementation of ALA vs. LA for
2 yr significantly reduced serum CRP but had no effect on other
inflammatory markers (39).
Our findings show that plasma levels of AA and omega-3
PUFAs, which probably reflect higher dietary intake, are asso-
ciated with lower serum concentrations of certain proinflam-
matory cytokines and lower concentrations of certain antiin-
flammatory cytokines. Therefore, these findings support the
view that AA and n-3 PUFAs may modulate the inflammatory
response by acting both on the proinflammatory and antiin-
flammatory arms of the cytokine network. Such a modulatory
effect on multiple signaling pathways suggests a direct regu-
latory effect on gene expression. Interestingly, short-term in-
fusion of n-3 lipid emulsion markedly suppresses monocytic
generation of TNF
␣
, IL-1, IL-6, and IL-8 in response to endo-
toxin (40).
The mechanism by which AA and n-3 fatty acids may
inhibit the production of proinflammatory cytokines has
been intensively investigated. Fatty acids can bind to the
peroxisome proliferator activated receptors (PPaR)
␣
and
PPaR
␥
, which regulate the transcription of target genes (7,
41). PPaRs can also repress gene transcription by interfering
with signaling molecules, such as nuclear factor-
B, there-
fore inhibiting the production of proinflammatory cytokines
(42, 43). De Caterina and colleagues (8, 44) found that poly-
unsaturated fatty acids have antiinflammatory properties
and hypothesized that they exert this effect because they
have an unsaturated double bond, which, regardless of the
n-3 or n-6 position, inactivates reactive oxygen species and
prevents their interaction with nuclear factor-
B. This hy-
pothesis is consistent with our findings suggesting that n-3
and n-6 fatty acids have both antiinflammatory properties.
TABLE 3. Multivariate analysis of the relationship between total n-3 and n-6 fatty acids and inflammatory markers
Total n-3 fatty acids,
% (quartiles)
a
(mean ⫾ SD, 3.4 ⫾ 1.0)
P for
trend
Total n-6 fatty acids,
% (quartiles)
a
(mean ⫾ SD, 33.1 ⫾ 4.6)
P for
trend
n-6/n-3 fatty acids,
% (quartiles)
(mean ⫾
SD, 10.5 ⫾ 3.1)
P for
trend
Quartiles limits ⬍2.7 2.7–3.3 3.4–3.9 ⬎3.9
ref
⬍30 30–33 34–36.3 ⬎36
ref
⬍8.4–10.2 10.2–12.1 ⬎12.1
ref
Median 2.3 3.0 3.5 4.5 28.0 31.6 34.7 38.2 7.3 9.3 11.0 13.8
IL-6 (pg/ml) 1.38
b
1.29 1.24
c
1.14 0.0048 1.26 1.29 1.29 1.20 0.56 1.13
b
1.29 1.24 1.39 0.0057
sIL-6r (ng/ml) 79
b
84
b
91 97 0.0003 88 85 90 88 0.74 89 87 88 86 0.53
IL-1

(pg/ml) 0.13 0.14 0.13 0.13 0.53 0.13 0.13 0.14 0.13 0.56 0.13 0.13 0.13 0.14 0.93
IL-1ra (pg/ml) 140
b
131 130 121 0.0044 143 128 128 123 0.0258 122
b
126
c
132 142 0.0007
TNF-
␣
(pg/ml) 5.1
c
4.6 4.9 4.3 0.0404 4.8 4.9 4.6 4.5 0.32 4.3 4.7 5.0 4.9 0.08
IL-10 (pg/ml) 2.6 4.1 4.2 4.7 0.0236 4.2 4.0 3.1 3.3 0.07 4.5
b
4.0
a
3.9 2.9 0.0002
TGF-

(ng/ml) 11.1
b
11.4
b
12.0 13.1 0.0012 10.6 12.0 12.3 12.7 0.014 13.1 12.1 11.2 11.0 0.0005
CRP (mg/liter) 2.51 2.40 2.54 2.24 0.35 2.53 2.28 2.55 2.32 0.77 2.30 2.45 2.45 2.48 0.46
ref, Reference quartile.
a
Mean values and statistics for the association of n-3 and n-6 fatty acids with selected inflammatory markers are mutually independent. All mean values and statistical tests
are adjusted for age, sex, education, daily intake of energy, proteins, lipids and carbohydrates, physical activity, BMI, smoking, LDL cholesterol, HDL cholesterol, triglycerides,
hypertension, diabetes, coronary heart disease, congestive heart failure, stroke, peripheral artery disease, and potentially confounding drug treatment.
b
P ⬍ 0.01, compared with the reference quartile.
c
P ⬍ 0.05, compared with the reference quartile.
444 J Clin Endocrinol Metab, February 2006, 91(2):439–446 Ferrucci et al. • PUFA Levels Correlate with Low Inflammation
on February 24, 2006 jcem.endojournals.orgDownloaded from
The most important limitation of this study is the cross-
sectional nature of our analysis. Although the consistency of the
effect across multiple inflammatory markers is suggestive of
causality, the correlation reported in this study does not prove
the link between PUFAs and inflammatory markers but sug-
gests that the physiological concentration of PUFAs reflects the
severity of inflammation independently of other risk factors. In
addition, although our analysis was adjusted for a number of
potential confounders, we cannot exclude the possibility that
other factors affect both n-3 fatty acids and cytokine concen-
trations. The diet of the Tuscany population is particularly poor
of polyunsaturated fatty acids (45). In the InCHIANTI study,
the average estimated daily intake of PUFAs was 7.4 g, which
is lower than the intake reported for other populations (46) and
even compared with other Italian populations (47). On the
contrary, the intake of monounsaturated fatty acids in our pop-
ulation tended to be high (⬎50% of total lipids), likely the result
of the large consumption of olive oil in Italy and, in particular,
in the Tuscany region. The generalizability of our findings to
other populations with different dietary intake should be con-
firmed by other studies.
Our finding that n-3 and n-6 account for a significantly
lower percentage of total fatty acids in older persons may
explain the mild proinflammatory state that is often found in
the elderly and is not completely accounted for by cardio-
vascular risk factors and morbidity (48). If this hypothesis is
correct, nutritional intervention may contrast this age-related
trend to a proinflammatory state.
This study also has several strengths. To our knowledge,
this is the first investigation of the relationship between
plasma concentrations of fatty acids and multiple proinflam-
matory and antiinflammatory cytokines based on a repre-
sentative sample of the general population. Different fatty
acids were directly measured in plasma and not estimated
from dietary reports. Because no participants were using
dietary supplements, our findings are based on physiological
plasma concentrations; this information is more precise and
provides a more objective measure of fatty acid exposure,
which depends on both intake and metabolism. Finally, in-
formation on multiple potential confounders, including di-
etary intake, was available for all participants.
Because serum levels of specific fatty acids can be easily
modified by a different selection of foods in the diet or dietary
supplementation, physicians should consider dietary interven-
tions to suppress production of proinflammatory compounds
as part of the prevention and treatment of diseases in which
inflammation exerts adverse effects on clinical progression.
Acknowledgments
The authors are indebted to Chiara Ceccucci for integrating fatty acids
peaks in the chromatograms and for data entry.
Received June 16, 2005. Accepted October 7, 2005.
Address all correspondence and requests for reprints to: Luigi Fer-
rucci, M.D., Ph.D., Longitudinal Studies Section, Clinical Research
Branch, National Institute on Aging, National Institutes of Health, Har-
bor Hospital, 5th Floor, 3001 Hanover Street, Baltimore, Maryland 21225.
E-mail: ferruccilu@grc.nia.nih.gov.
This work was supported as a targeted project by the Italian Ministry
of Health (ICS110.1/R597.71) and an unrestricted grant by Bracco Im-
aging SpA, Italy (to A.Ch.). C.A.-L. was supported by the Ramon Cajal
program by the Ministry of Science and Technology, Spain.
References
1. Albert CM, Campos H, Stampfer MJ, Ridker PM, Manson JE, Willett WC,
Ma J 2002 Blood levels of long-chain n-3 fatty acids and the risk of sudden
death. N Engl J Med 346:1113–1118
2. Ascherio A, Rimm EB, Stampfer MJ, Giovannucci EL, Willett WC 1995
Dietary intake of marine n-3 fatty acids, fish intake, and the risk of coronary
disease among men. N Engl J Med 332:977–982
3. Bucher HC, Hengstler P, Schindler C, Meier G 2002 N-3 polyunsaturated fatty
acids in coronary heart disease: a meta-analysis of randomized controlled
trials. Am J Med 112:298–304
4. Kromhout D, Bosschieter EB, de Lezenne Coulander C 1985 The inverse
relation between fish consumption and 20-year mortality from coronary heart
disease. N Engl J Med 312:1205–1209
5. Prichard BN, Smith CC, Ling KL, Betteridge DJ 1995 Fish oils and cardio-
vascular disease. BMJ 310:819 –820
6. Calder PC 2002 Dietary modification of inflammation with lipids. Proc Nutr
Soc 61:345–358
7. Khan SA, Vanden Heuvel JP 2003 Role of nuclear receptors in the regulation
of gene expression by dietary fatty acids. J Nutr Biochem 14:554–567 (Review)
8. De Caterina R, Libby P 1996 Control of endothelial leukocyte adhesion mol-
ecules by fatty acids. Lipids 31(Suppl):S57–S63
9. Calder PC 2003 N-3 polyunsaturated fatty acids and inflammation: from
molecular biology to the clinic. Lipids 38:343–352
10. Endres S, Meydani SN, Ghorbani R, Schindler R, Dinarello CA 1993 Dietary
supplementation with n-3 fatty acids suppresses interleukin-2 production and
mononuclear cell proliferation. J Leukoc Biol 54:599 – 603
11. Mantzioris E, Cleland LG, Gibson RA, Neumann MA, Demasi M, James MJ
2000 Biochemical effects of a diet containing foods enriched with n-3 fatty
acids. Am J Clin Nutr 72:42– 48
12. Geelen A, Brouwer IA, Schouten EG, Kluft C, Katan MB, Zock PL 2004 Intake
of n-3 fatty acids from fish does not lower serum concentrations of C-reactive
protein in healthy subjects. Eur J Clin Nutr 58:1440–1442
13. Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Willett WC, Rimm EB
2003 Habitual dietary intake of n-3 and n-6 fatty acids in relation to inflam-
matory markers among U.S. men and women. Circulation 108:155–160
14. Seierstad SL, Seljeflot I, Johansen O, Hansen R, Haugen M, Rosenlund G,
Froyland L, Arnesen H 2005 Dietary intake of differently fed salmon; the
influence on markers of human atherosclerosis. Eur J Clin Invest 35:52–59
15. Fusconi E, Pala V, Riboli E, Vineis P, Sacerdote C, Del Pezzo M, Santucci de
Magistris M, Palli D, Masala G, Sieri S, Foggetti CE, Giurdanella MC,
Tumino R, Krogh V 2003 Relationship between plasma fatty acid composition
and diet over previous years in the Italian centers of the European Prospective
Investigation into Cancer and Nutrition (EPIC). Tumori 89:624 – 635
16. Ferrucci L, Bandinelli S, Benvenuti E, Di Iorio A, Macchi C, Harris TB,
Guralnik JM 2000 Subsystems contributing to the decline in ability to walk:
bridging the gap between epidemiology and geriatric practice in the InCHI-
ANTI study. J Am Geriatr Soc 48:1618 –1625
17. Poehlman ET, Dvorak RV 2000 Energy expenditure, energy intake, and
weight loss in Alzheimer disease. Am J Clin Nutr 71:650S–655S
18. American Psychiatric Association 1987 Diagnostic and statistical manual of
mental disorders. 3rd ed. rev. (DSM-III-R). Washington, DC: American Psy-
chiatric Association
19. Folch J, Lees M, Sloane Stanley GH 1957 A simple method for the isolation
and purification of total lipides from animal tissues. J Biol Chem 226:497–509
20. Lepage G, Roy CC 1984 Improved recovery of fatty acid through direct trans-
esterification without prior extraction or purification. J Lipid Res 25:1391–1396
21. Rodriguez-Palmero M, Lopez-Sabater MC, Castellote-Bargallo AI, De la
Torre-Boronat MC, Rivero-Urgell M 1998 Comparison of two methods for the
determination of fatty acid profiles in plasma and erythrocytes. J Chromatogr
A 793:435– 440
22. Kang JX 2003 The importance of omega-6/omega-3 fatty acid ratio in cell
function. The gene transfer of omega-3 fatty acid desaturase. World Rev Nutr
Diet 92:23–36
23. Pisani P, Faggiano F, Krogh V, Palli D, Vineis P, Berrino F 1997 Relative
validity and reproducibility of a food frequency dietary questionnaire for use
in the Italian EPIC centres. Int J Epidemiol 26(Suppl 1):S152–S160
24. Guralnik JM, Fried LP, Simonsick EM, Kasper JD, Lafferty ME 1995 The
Women’s Health and Aging Study. Health and social characteristics of older
women with disability. Bethesda, MD: National Institute of Aging
25. McDermott MM, Greenland P, Liu K, Guralnik JM, Celic L, Criqui MH,
Chan C, Martin GJ, Schneider J, Pearce WH, Taylor LM, Clark E 2002 The
ankle brachial index is associated with leg function and physical activity: the
Walking and Leg Circulation Study. Ann Intern Med 136:873–883
26. Yeh LL, Kuller LH, Bunker CH, Ukoli FA, Huston SL, Terrell DF 1996 The role
of socioeconomic status and serum fatty acids in the relationship between intake
of animal foods and cardiovascular risk factors. Ann Epidemiol 6:290 –298
27. Sera RK, McBride JH, Higgins SA, Rodgerson DO 1994 Evaluation of ref-
erence ranges for fatty acids in serum. J Clin Lab Anal 8:81– 85
Ferrucci et al. • PUFA Levels Correlate with Low Inflammation J Clin Endocrinol Metab, February 2006, 91(2):439–446 445
on February 24, 2006 jcem.endojournals.orgDownloaded from
28. James MJ, Proudman SM, Cleland LG 2003 Dietary n-3 fats as adjunctive
therapy in a prototypic inflammatory disease: issues and obstacles for use in
rheumatoid arthritis. Prostaglandins Leukot Essent Fatty Acids 68:399 – 405
29. Pestka S, Krause CD, Sarkar D, Walter MR, Shi Y, Fisher PB 2004 Interleu-
kin-10 and related cytokines and receptors. Annu Rev Immunol 22:929–979
30. Wahl SM, Chen W 2003 TGF

: how tolerant can it be? Immunol Res 28:167–179
31. Dinarello CA 2000 The role of the interleukin-1-receptor antagonist in blocking
inflammation mediated by interleukin-1. N Engl J Med 343:732–734
32. Arend WP 2002 The balance between IL-1 and IL-1Ra in disease. Cytokine
Growth Factor Rev 13:323–340
33. Biasucci LM, Liuzzo G, Fantuzzi G, Caligiuri G, Rebuzzi AG, Ginnetti F,
Dinarello CA, Maseri A 1999 Increasing levels of interleukin (IL)-1Ra and IL-6
during the first 2 days of hospitalization in unstable angina are associated with
increased risk of in-hospital coronary events. Circulation 99:2079 –2084
34. Gabay C, Smith MF, Eidlen D, Arend WP 1997 Interleukin 1 receptor an-
tagonist (IL-1Ra) is an acute-phase protein. J Clin Invest 99:2930–2940
35. Schroers A, Hecht O, Kallen KJ, Pachta M, Rose-John S, Grotzinger J 2005
Dynamics of the gp130 cytokine complex: a model for assembly on the cellular
membrane. Protein Sci 14:783–790
36. Taga T, Hibi M, Hirata Y, Yamasaki K, Yasukawa K, Matsuda T, Hirano T,
Kishimoto T 1989 Interleukin-6 triggers the association of its receptor with a
possible signal transducer, gp130. Cell 58:573–581
37. Abbatecola AM, Ferrucci L, Grella R, Bandinelli S, Bonafe M, Barbieri M,
Corsi AM, Lauretani F, Franceschi C, Paolisso G 2004 Diverse effect of in-
flammatory markers on insulin resistance and insulin-resistance syndrome in
the elderly. J Am Geriatr Soc 52:399 – 404
38. Gabay C, Kushner I 1999 Acute-phase proteins and other systemic responses
to inflammation. N Engl J Med 340:448 – 454
39. Bemelmans WJ, Lefrandt JD, Feskens EJ, van Haelst PL, Broer J, Mey-
boom-de Jong B, May JF, Tervaert JW, Smit AJ 2004 Increased
␣
-linolenic acid
intake lowers C-reactive protein, but has no effect on markers of atheroscle-
rosis. Eur J Clin Nutr 58:1083–1089
40. Mayer K, Meyer S, Reinholz-Muhly M, Maus U, Merfels M, Lohmeyer J,
Grimminger F, Seeger W 2003 Short-time infusion of fish oil-based lipid
emulsions, approved for parenteral nutrition, reduces monocyte proinflam-
matory cytokine generation and adhesive interaction with endothelium in
humans. J Immunol 171:4837– 4843
41. Jump DB 2002 The biochemistry of n-3 polyunsaturated fatty acids. J Biol
Chem 277:8755– 8758
42. Marx N, Sukhova GK, Collins T, Libby P, Plutzky J 1999 PPAR
␣
activators
inhibit cytokine-induced vascular cell adhesion molecule-1 expression in hu-
man endothelial cells. Circulation 99:3125–3131
43. Staels B, Koenig W, Habib A, Merval R, Lebret M, Torra IP, Delerive P, Fadel
A, Chinetti G, Fruchart JC, Najib J, Maclouf J, Tedgui A 1998 Activation of
human aortic smooth-muscle cells is inhibited by PPAR
␣
but not by PPAR
␥
activators. Nature 393:790 –793
44. De Caterina R, Spiecker M, Solaini G, Basta G, Bosetti F, Libby P, Liao J 1999
The inhibition of endothelial activation by unsaturated fatty acids. Lipids
34(Suppl):S191–S194
45. Sofi F, Vecchio S, Giuliani G, Marcucci R, Gori AM, Fedi S, Casini A,
Surrenti C, Abbate R, Gensini GF 2005 Dietary habits, lifestyle and cardio-
vascular risk factors in a clinically healthy Italian population: the ‘Florence’ diet
is not Mediterranean. Eur J Clin Nutr 59:584 –591
46. Hakala P, Knuts LR, Vuorinen A, Hammar N, Becker W 2003 Comparison of
nutrient intake data calculated on the basis of two different databases. Results
and experiences from a Swedish-Finnish study. Eur J Clin Nutr 57:1035–1044
47. Saba A, Turrini A, Mistura G, Cialfa E 1990 Indagine nazionale sui consumi
alimentari delle famiglie 1980 –1984 —alcuni principali risultati. L Riv Ital Sci
Alimentaziano 19:453– 465
48. Ferrucci L, Corsi A, Lauretani F, Bandinelli S, Bartali B, Taub DD, Guralnik
JM, Longo DL 2005 The origins of age-related proinflammatory state. Blood
105:2294–2299
JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the
endocrine community.
446 J Clin Endocrinol Metab, February 2006, 91(2):439–446 Ferrucci et al. • PUFA Levels Correlate with Low Inflammation
on February 24, 2006 jcem.endojournals.orgDownloaded from
