1. Plasma n-3 and n-6 fatty acids and the incidence of
atrial fibrillation following coronary artery bypass
graft surgery
Gudrun V. Skuladottir*,†
, Ragnhildur Heidarsdottir*,†
, David O. Arnar†,‡
, Bjarni Torfason†,§
, Vidar Edvardsson†,¶
,
Gizur Gottskalksson‡
, Runolfur Palsson†,**
and Olafur S. Indridason**
*
Department of Physiology, †
Faculty of Medicine, School of Health Sciences, University of Iceland, ‡
Division of Cardiology,
Department of Medicine, Cardiovascular Research Center, §
Department of Cardiothoracic Surgery, ¶
Children’s Medical
Center, **
Division of Nephrology, Department of Medicine, Landspitali – The National University Hospital of Iceland,
Hringbraut, Reykjavik, Iceland
ABSTRACT
Background The anti-inflammatory or anti-arrhythmic effects of n-3 long-chain polyunsaturated fatty acids
(LC-PUFA) may decrease the risk of postoperative atrial fibrillation (POAF), but interventional studies have
yielded conflicting results. We examined the association between n-3 LC-PUFA and n-6 LC-PUFA in plasma
phospholipids (PL) and POAF in patients undergoing coronary artery bypass grafting (CABG).
Methods A total of 125 patients undergoing CABG were enrolled in the study. The levels of fatty acids in PL
were measured preoperatively and on the third postoperative day. The endpoint was defined as POAF lasting
‡5 min. The incidence of POAF was compared between quartiles of the level of each fatty acid in plasma PL by
univariate and multivariable analysis.
Results The incidence of POAF was 49Æ6%. By univariate analysis, the incidence of POAF increased
significantly with each higher quartile of pre- and postoperative docosahexaenoic acid (DHA) and diminished
significantly with each higher quartile of pre- and postoperative arachidonic acid (AA). For postoperative total n-3
LC-PUFA, there was a significant U-curve relationship where the second quartile had the lowest incidence of
POAF or 25Æ8%. In multivariable analysis, this U-curve relationship between n-3 LC-PUFA levels and POAF risk
was not significant, whereas the association between POAF and DHA or AA remained statistically significant.
Conclusions This study suggests that n-3 LC-PUFA supplements might prevent POAF in CABG patients with
low baseline levels of these fatty acids in plasma PL, but may be harmful in those with high levels. AA may play
an important role in electrophysiological processes.
Keywords Atrial fibrillation, coronary artery bypass grafting, C-reactive protein, n-3 long-chain polyunsaturated
fatty acids, postoperative, preoperative.
Eur J Clin Invest 2011
Introduction
Atrial fibrillation occurs commonly after open-heart surgery
[1]. While advanced age and the increased complexity of the
surgical procedure are both major risk factors for postoperative
atrial fibrillation (POAF) [2], the pathogenesis of the disorder is
largely unknown. Both neurohumoral activation and a robust
inflammatory response after surgery have been implicated in
the development of this arrhythmia [3–6].
n-3 Long-chain polyunsaturated fatty acids (LC-PUFA) have
been viewed as potential preventive therapeutic options in
POAF, because of their anti-inflammatory and ⁄ or anti-arrhyth-
mic action [7,8]. However, the results of recent interventional
studies investigating the effect of n-3 LC-PUFA supplementa-
tion on the incidence of POAF following open-heart surgery
have yielded conflicting results [9–12]. Neither of the two stud-
ies that showed beneficial effects of n-3 LC-PUFA on POAF fol-
lowing coronary artery bypass graft surgery (CABG) included
analysis of n-3 LC-PUFA levels in blood phospholipids (PL)
[9,10]. These studies were conducted in populations expected
to have low baseline levels of n-3 LC-PUFA in plasma PL and
high levels of n-6 LC-PUFA [13]. In contrast, our previous study
European Journal of Clinical Investigation 1
DOI: 10.1111/j.1365-2362.2011.02497.x
ORIGINAL ARTICLE

2. [11] and a study by other investigators [12] which were carried
out in populations with relatively high baseline levels of n-3
LC-PUFA in plasma PL showed no effect of n-3 LC-PUFA on
AF incidence following open-heart surgery. Thus, n-3 LC-
PUFA treatment may only be effective in the prevention of
POAF in patients with low baseline levels of these fatty acids in
blood PL.
In this study, we examined the association between levels of
n-3 LC-PUFA and n-6 LC-PUFA in plasma PL and the
incidence of POAF in patients undergoing CABG.
Materials and methods
Subjects and study design
This study was a part of a larger prospective, randomized, dou-
ble-blinded, placebo-controlled clinical trial on the use of n-3
LC-PUFA for a week prior to open-heart surgery to prevent
POAF. Details of the study have been presented previously
[11]. In brief, all patients scheduled for elective or semi-emer-
gent open heart surgery between August 2007 and May 2009
were evaluated for participation, and a total of 170 patients
were enrolled. The treatment group received 2240 mg of n-3
LC-PUFA daily, providing 1240 mg eicosapentaenoic acid
(EPA) and 1000 mg docosahexaenoic acid (DHA) as ethyl
esters. The n-3 LC-PUFA was administered orally in capsules
that are commercially available in Iceland (Omega Forte; Lysi
Inc, Reykjavı´k, Iceland). The placebo treatment consisted of
2000 mg olive oil daily. Exclusion criteria included age < 40, a
history of supraventricular arrhythmias or the current use of
the anti-arrhythmic medications amiodarone and ⁄ or sotalol. To
eliminate confounding resulting from different complexity of
the surgical procedure, subjects undergoing open-heart proce-
dures other than CABG were excluded from this analysis. The
patients answered a questionnaire on lifestyle issues, including
the frequency of fish consumption but not the type of fish
consumed.
All participants received standard care following the surgical
procedure, and all had continuous electrocardiographic moni-
toring while being hospitalized. The study endpoint, POAF,
was defined as an episode lasting more than 5 min. All patients
gave written informed consent. The study was approved by the
Bioethics Committee of Landspitali – The National University
Hospital of Iceland, and the Icelandic Data Protection Author-
ity. Reporting of the study conforms to the STROBE statement
along with references to STROBE and the broader EQUATOR
guidelines [14].
Measurement of CRP and fatty acids in plasma
phospholipids
Venous blood samples were obtained immediately before sur-
gery (preoperative) and on the third postoperative day. Plasma
was separated from whole blood by immediate centrifugation
at 1000 g for 10 min, and the samples were frozen at )76 °C and
stored until the analysis of the fatty acid levels in PL was
performed.
Serum levels of C-reactive protein (CRP) were measured by
an enzymatic sandwich immunoassay (Vitros 5.1 FS Chemistry
System; Ortho-Clinical Diagnostics, Raritan, NJ, USA) prior to
surgery and daily during the postoperative period until the
levels had peaked.
For the analysis of fatty acids in plasma PL, the total plasma
lipid fraction was first extracted with chloroform : methanol
(2 : 1, v ⁄ v), using the method of Folch et al. [15]. The antioxi-
dant butylated hydroxytoluene (BHT, 5 mg per 100 mL) was
added to the extraction medium. The PL were separated on a
thin-layer chromatography plate (Adsorbosil H, Alltech, Deer-
field, IL, USA) using the solvent system petroleum ether ⁄ dieth-
ylether ⁄ acetic acid (80 : 20 : 1, v ⁄ v ⁄ v). The PL fatty acids were
methylated with 14% boron trifluoride ⁄ methanol (Sigma
Chemical Co., St. Louis, MO, USA) for 45 min at 110 °C. The
fatty acid methyl esters were analysed in a HP Series II 5890 A
Gas Chromatograph (Hewlett Packard Co ⁄ Agilent, Palo Alto,
CA, USA) equipped with a flame ionization detector and a
Chrompack CP-WAX 52CB capillary column (25 m · 0Æ32 mm
i.d. · 0Æ2 lm film thickness). The oven was programmed to pro-
vide an initial temperature of 90 °C for 2 min, then increasing
temperature by 30 °C per min to 165 °C and finally by 3 °C per
min to 225 °C. The carrier gas was hydrogen. The fatty acid
methyl esters were identified and calibrated against commer-
cial standards (Sigma Chemical Co.; Nu-Check-Prep, Elysian,
MN, USA). The software HP 3365 Chemstation, Version AÆ02Æ12
(Agilent, Palo Alto, CA, USA), was used for instrumental con-
trol and data acquisition and processing. The results are
expressed as percentage (%) of total fatty acids in plasma PL.
Statistical analysis
As there was no difference in the incidence of POAF between
the n-3 PUFA and placebo groups [11], the treatment assign-
ment was ignored in this analysis. Differences in baseline and
operative characteristics between the patients with POAF
(POAF group) and those without POAF (non-POAF group)
were compared with Wilcoxon–Mann–Whitney test for contin-
uous variables and the chi squared test or Fisher’s exact test for
dichotomous variables. An independent samples t-test was
used to compare the difference in fatty acid composition of
plasma PL between patients with POAF and those without
POAF and a paired t-test to compare changes within groups.
To examine the association between the levels of individual
fatty acids and POAF, we compared the rate of POAF between
quartiles of the fatty acid levels using chi squared test and the
Somers’d for ordinal variables. To examine independent associ-
ation between the quartiles of fatty acid levels and POAF, a
2 ª 2011 The Authors. European Journal of Clinical Investigation ª 2011 Stichting European Society for Clinical Investigation Journal Foundation
G. V. SKULADOTTIR ET AL. www.ejci-online.com

3. logistic regression analysis was used with POAF as the depen-
dent variable and the quartiles of each fatty acid as a categorical
variable, adjusting for age, BMI, smoking and peak postopera-
tive CRP. Data are presented as median (range), percentages or
mean ± standard error of mean (SEM), unless otherwise noted.
Two-sided P value < 0Æ05 was considered statistically signifi-
cant. All statistical analyses were carried out using SPSS
software (version 11Æ0, IBM Corporation, Somers, NY, USA).
Results
A total of 125 patients underwent CABG, of whom 62 patients
(49Æ6%) developed POAF. Six patients with missing data on
fatty acid levels, four preoperatively and two postoperatively,
were not included in the respective analyses. The baseline and
operative characteristics of the POAF group compared with the
non-POAF group are shown in Table 1. The POAF group was
older (P = 0Æ003) and their body mass index (BMI) was lower
(P = 0Æ026) compared with the non-POAF group. The plasma
CRP levels peaked on the third postoperative day (median,
range, day 1–10). The median peak CRP level was higher in the
POAF group compared with the non-POAF group, 218Æ0
(66Æ0–492Æ0) mg L)1
vs. 201Æ0 (34Æ0–370Æ0) mg L)1
, respectively
(P < 0Æ05).
The preoperative plasma PL levels of fatty acids in the two
groups are shown in Fig. 1. The POAF group had lower levels
of arachidonic acid (AA) (P < 0Æ05) and higher levels of DHA
Table 1 Baseline and operative characteristics of patients who did (POAF group) or did not (non-POAF group) develop
postoperative atrial fibrillation (POAF)
Non-POAF group (n = 63) POAF group (n = 62) P value
Age (years) 66 (45–79) 69 (45–82)* 0Æ003
BMI (kg m)2
) 28Æ4 (20Æ9–41Æ3) 27Æ0 (19Æ1–35Æ0)* 0Æ026
Gender (% men) 84Æ1 79Æ0 0Æ462
Smoking (%) 27Æ0 14Æ5 0Æ086
Fish intake (%, >once a week) 62Æ3 82Æ3* 0Æ013
Cod liver oil intake (%) 54Æ0 56Æ5 0Æ367
n-3 LC-PUFA intake (%) 25Æ4 27Æ4 0Æ364
Use of b blockers (%) 84Æ1 77Æ4 0Æ341
Use of statins (%) 90Æ5 83Æ9 0Æ27
Hypertension (%) 65Æ1 64Æ5 0Æ947
Diabetes (%) 19Æ0 12Æ9 0Æ349
Estimated operative blood loss (mL) 900 (0–2860) 800 (0–6200) 0Æ799
Blood volume in drains (mL) 700 (110–3070) 775 (96–4980) 0Æ664
Peak postoperative CRP level (mg L)1
) 197Æ0 (34Æ0–370Æ0) 216Æ5 (36Æ0–416Æ0)* 0Æ042
ECC time (min) 84Æ5 (0–180) 80 (0–183) 0Æ696
Off pump surgery (%) 19Æ0 12Æ9 0Æ349
Aortic cross-clamp time (min) 45 (0–87) 42 (0–120) 0Æ765
Time in intensive care unit (hours) 21Æ5 (4–209Æ5) 22 (14Æ5–213) 0Æ33
Need for inotropic support (n)
None or minor 52 48 0Æ64
Short term 9 10
Long term 2 4
Data are expressed as median (range), percentage or number of subjects. BMI, body mass index; CRP, C-reactive protein; LC-PUFA, long-chain polyunsaturated
fatty acids; ECC, extracorporeal circulation.
*P < 0Æ05, compared with the non-POAF group, Wilcoxon–Mann–Whitney or chi-squared tests.
European Journal of Clinical Investigation 3
PLASMA N-3 AND N-6 FATTY ACIDS AND POSTOPERATIVE ATRIAL FIBRILLATION

4. (P < 0Æ05) compared with the non-POAF group. On the third
postoperative day, the plasma PL levels of AA, EPA and total
n-3 LC-PUFA were lower than preoperative levels (P < 0Æ05),
while those of DHA remained unchanged in both the non-
POAF (Fig. 2a) and POAF groups (Fig. 2b). However, the level
of AA was still significantly lower and the levels of EPA and
DHA were higher in the POAF group compared with the non-
POAF group (7Æ94 ± 0Æ24% vs. 8Æ99 ± 0Æ28%, 2Æ77 ± 0Æ17% vs.
2Æ33 ± 0Æ14%, and 6Æ93 ± 0Æ15% vs. 6Æ40 ± 0Æ14%, respectively,
P < 0Æ05).
Table 2 shows the incidence of POAF according to quar-
tiles of pre- and postoperative AA, EPA, DHA and total n-3
LC-PUFA levels in plasma PL. There was a significant differ-
ence in POAF incidence between quartiles of both pre- and
postoperative AA levels (P = 0Æ05 and P = 0Æ006, respec-
tively), with a significant trend for a decreasing incidence of
POAF with higher quartiles (P = 0Æ008 and P = 0Æ003, respec-
tively). The incidence of POAF was also significantly differ-
ent between quartiles of preoperative DHA levels (P = 0Æ023),
but in the case of postoperative DHA levels, the difference
did not quite reach statistical significance (P = 0Æ068). There
was, however, a significant trend for an increasing incidence
of POAF with higher quartiles of both pre- and postoperative
DHA levels (P = 0Æ003 and P = 0Æ008, respectively). No signif-
icant difference in POAF incidence was observed between
the quartiles of preoperative total n-3 LC-PUFA levels
(P = 0Æ41), nor was there a significant trend for an increase in
incidence (P = 0Æ13). Postoperatively, a nonsignificant trend
toward an increase in the POAF incidence between the
quartiles of total n-3 LC-PUFA levels was observed
(P = 0Æ055). However, the association between the total n-3
LC-PUFA levels and the POAF incidence was significant
(P = 0Æ014), with the second quartile of total n-3 LC-PUFA
levels having the lowest POAF rate (25Æ8%), significantly
lower than the first (P = 0Æ05), third (P = 0Æ02) and fourth
(P = 0Æ002) quartiles (Fig. 3). The trend was similar for EPA
albeit nonsignificant.
There was a significant direct association between age and
the levels of total n-3 LC-PUFA and an inverse association with
the levels of AA. Moreover, those who reported use of cod liver
oil supplementation and more frequent fish consumption more
than once per week were significantly older than those who did
not. In the subsequent logistic regression analysis, adjustments
were therefore made for age, as well as BMI, smoking and max-
imal peak postoperative CRP. Figure 4 shows the adjusted odds
ratio of POAF for each quartile of the fatty acids where the
Figure 1 Preoperative plasma phospholipid levels (% of total
fatty acids) of arachidonic acid (AA, 20:4n-6), eicosapentaenoic
acid (EPA, 20:5n-3), docosahexaenoic acid (DHA, 22:6n-3), and
total n-3 long-chain polyunsaturated fatty acids (LC-PUFA) in
the nonpostoperative atrial fibrillation (non-POAF) group and
POAF group. Values are mean ± SEM. *P < 0Æ05, compared
with the non-POAF group, independent samples t-test.
Figure 2 Plasma phospholipid levels (% of total fatty acids) of
arachidonic acid (AA, 20:4n-6), eicosapentaenoic acid (EPA,
20:5n-3), docosahexaenoic acid (DHA, 22:6n-3) and total n-3
long-chain polyunsaturated fatty acids (LC-PUFA) in the non-
postoperative atrial fibrillation (non-POAF) group (a) and POAF
group (b), immediately before surgery (preoperative) and on
the third postoperative day. Values are mean ± SEM. *P < 0Æ05,
compared with the preoperative values, paired samples t-test.
4 ª 2011 The Authors. European Journal of Clinical Investigation ª 2011 Stichting European Society for Clinical Investigation Journal Foundation
G. V. SKULADOTTIR ET AL. www.ejci-online.com

5. lowest quartile is the reference. The patients with preoperative
AA levels in the third quartile had significantly lower odds
ratio of POAF (OR 0Æ246, 95% CI 0Æ071–0Æ845) (Fig. 4a), and
postoperatively, the patients in the second (OR 0Æ260, 95% CI
0Æ076–0Æ886) and fourth quartiles (OR 0Æ145, 95% CI 0Æ040–0Æ531)
had significantly lower odds ratio of POAF (Fig. 4b) than
patients who had levels in the reference quartiles. Those in the
highest quartile of both pre- and postoperative DHA levels had
a significantly higher odds ratio of POAF than the lowest one,
with OR of 3Æ358 (95% CI 1Æ026–10Æ995) (Fig. 4e) and 4Æ902 (95%
CI 1Æ395–17Æ222) (Fig. 4f), respectively. For the postoperative
total n-3 LC-PUFA level, there was a nonsignificant U-curve
association with POAF with patients in the second quartile hav-
ing a lower odds ratio of POAF (OR 0Æ452, 95% CI 0Æ132–1Æ550)
than those in the first, third and fourth quartiles, but this
difference was not significant (Fig. 4h).
Discussion
In this prospective cohort study of patients undergoing CABG
surgery, we found that lower levels of DHA and higher levels
of AA in plasma PL were associated with decreased risk of
POAF. For total n-3 LC-PUFA levels, there was a trend toward
a U-curve relationship, suggesting that these fatty acids may be
of benefit in populations with low intake of n-3 LC-PUFA.
Our findings are somewhat surprising and contradict the
widely held belief that the anti-inflammatory and proposed
anti-arrhythmic effects of n-3 LC-PUFA may be of value in
preventing POAF and other arrhythmias [16,17]. Indeed, two
recent open label studies demonstrated that short-term EPA
and DHA administration might be promising as a therapy for
prevention of AF following CABG surgery [9,10]. One of these
studies performed by Calo and coworkers in Italy [9] showed
Table 2 Incidence of postoperative atrial fibrillation (POAF) according to quartiles of arachidonic acid and n-3 LC-PUFA levels in
plasma phospholipids*
Quartiles P value
1 2 3 4 v2†
Somers’d‡
Preoperative (n = 30) (n = 30) (n = 31) (n = 30)
AA level (%)§
5Æ63–7Æ54 7Æ57–8Æ70 8Æ75–10Æ73 10Æ87–15Æ42
Incidence (%) 70Æ0 46Æ7 41Æ9 36Æ7 0Æ050 0Æ008
EPA level (%) 0Æ91–2Æ13 2Æ18–3Æ15 3Æ20–4Æ58 4Æ64–9Æ03
Incidence (%) 46Æ7 46Æ7 45Æ2 56Æ7 0Æ80 0Æ47
DHA level (%) 3Æ96–5Æ76 5Æ78–6Æ54 6Æ60–7Æ53 7Æ55–9Æ85
Incidence (%) 36Æ7 33Æ3 58Æ1 66Æ7 0Æ023 0Æ003
n-3 LC-PUFA level (%) 5Æ60–8Æ90 8Æ94–10Æ91 10Æ94–13Æ22 13Æ27–18Æ74
Incidence (%) 43Æ3 40Æ0 51Æ6 60Æ0 0Æ41 0Æ13
Postoperative (n = 30) (n = 31) (n = 31) (n = 31)
AA level (%) 4Æ97–6Æ92 6Æ96–8Æ17 8Æ18–9Æ53 9Æ57–16Æ57
Incidence (%) 74Æ2 40Æ0 51Æ6 32Æ3 0Æ006 0Æ003
EPA level (%) 0Æ89–1Æ61 1Æ66–2Æ17 2Æ24–3Æ36 3Æ39–6Æ12
Incidence (%) 46Æ7 38Æ7 45Æ2 67Æ7 0Æ12 0Æ078
DHA level (%) 4Æ24–5Æ81 5Æ87–6Æ54 6Æ61–7Æ54 7Æ58–9Æ48
Incidence (%) 36Æ7 38Æ7 58Æ1 64Æ5 0Æ068 0Æ008
n-3 LC-PUFA level (%) 6Æ00–8Æ61 8Æ68–9Æ87 9Æ94–11Æ56 11Æ66–16Æ62
Incidence (%) 50Æ0 25Æ8 58Æ1 64Æ5 0Æ014 0Æ055
*The n-6 Long-chain polyunsaturated fatty acid (LC-PUFA) arachidonic acid (AA); n-3 LC-PUFA include eicosapentaenoic acid (EPA), docosapentaenoic acid
(22:5n-3) and docosahexaenoic acid (DHA).
†
Pearson’s chi squared statistics.
‡
Somers’d statistics for trend in association between ordinal variables.
§
Levels are expressed as a percentage of total fatty acids.
European Journal of Clinical Investigation 5
PLASMA N-3 AND N-6 FATTY ACIDS AND POSTOPERATIVE ATRIAL FIBRILLATION

6. that the incidence of POAF in patients supplemented with
capsules containing EPA and DHA as ethyl esters for at least
5 days before surgery was 15% compared with 33% in the con-
trol group. In the other study, which was carried out by Heidt
and coworkers in Germany [10], a lower incidence of POAF
was observed in patients who received an intravenous infusion
of EPA and DHA as emulsion of highly refined fish oil, at least
12 h preoperatively and immediately following surgery com-
pared with the control group (17Æ3% vs. 30Æ6%, respectively).
The blood levels of the EPA and DHA were not assessed in
either of these studies. Both studies were conducted in popula-
tions reported to have low baseline levels of n-3 LC-PUFA in
plasma PL [13,18]. However, it is likely that the short-term
treatment with EPA and DHA resulted in increased levels of
these fatty acids in plasma PL. This suggestion is supported by
a study that showed an increase in the levels of EPA and DHA
in plasma PL within 8 h postprandially, when healthy volun-
teers received fish oil supplements [19]. The favourable effect of
n-3 LC-PUFA on POAF was recently contradicted by two dou-
ble-blind placebo-controlled studies, including our own study,
which found no benefit of short-term fish oil supplementation
on the risk of AF after CABG surgery [11,12]. Cod liver oil,
which is a rich source of EPA and DHA, is a commonly used
dietary supplement in Iceland, particularly among the elderly.
The cod liver oil consumption is reflected in relatively high
baseline plasma levels of these n-3 LC-PUFA in the Icelandic
population [20], including the patients in our study. The levels
of the n-3 LC-PUFA in serum PL were also relatively high at
entry in patients participating in the study carried out in the
United Kingdom [12]. Furthermore, n-3 LC-PUFA supplemen-
tation resulted in significantly higher levels of EPA and DHA
in PL in both serum [11,12] and right atrial tissue [12] compared
with the control group. These findings and our present observa-
tion of a U-curve relationship between the levels of total n-3
LC-PUFA in plasma PL and POAF suggest that n-3 LC-PUFA
administration may be beneficial for prevention of POAF in
patients with very low baseline levels of total n-3 LC-PUFA in
plasma PL. Moreover, our findings indicate that relatively high
levels of n-3 LC-PUFA in plasma PL, in particular DHA, appear
to have no protective effect and may even increase the risk of
POAF. The variable type of fish oil products used, the route
and ⁄ or the treatment period of administration might explain
the difference in the results of the aforementioned studies. Calo
and coworkers [9] administered fish oil through a nasogastric
tube in the immediate postoperative period, whereas Heidt and
coworkers [10] administered the oil intravenously. Improved
delivery of n-3 LC-PUFA during a critical part of the postopera-
tive period may be important for successful prevention of
POAF. In addition, different preparations may confer differen-
tial anti-arrhythmic effects or be incorporated into PL and cell
membranes in a variable fashion [21].
In the current study, we observed a novel inverse associa-
tion between AA levels in plasma PL and POAF. Earlier
studies have shown that the prostaglandin series derived
from AA have arrhythmic effects, which may be opposed by
free AA and EPA, indicating a distinction in the effect of the
fatty acids themselves and their metabolic products on car-
diac arrhythmia [17]. The human body contains excess of AA
as it is the most abundant dietary n-6 PUFA in the Western
diet along with its precursor linoleic acid [22]. In contrast, the
synthesis of EPA and DHA is limited in humans and, there-
fore, these fatty acids must come from the diet in order to
meet physiological needs [23]. It has been shown that dietary
n-3 LC-PUFA are incorporated into plasma PL, red blood
cells and myocardial tissue within 1 week of their ingestion,
causing a reduction in the AA levels in PL within a period of
1 month [24]. Thus, the balance between n-6 and n-3 LC-
PUFA in the diet determines their ratio in cell membranes
[25]. Our study showed the postoperative plasma PL levels of
AA and EPA, the source of n-6 and n-3-eicosanoids, to be
decreased compared to the preoperative levels, both in
patients who developed POAF and those who did not. This
finding may be explained by the increase in heparin-
enhanced plasma phospholipase A2 activity in patients
undergoing cardiac surgery [26]. In contrast, no change was
observed in the relatively high preoperative levels of DHA in
plasma PL in either group. A differential release of LC-PUFA
1
Quartiles of n-3 LC-PUFA
FrequencyofPOAF(%)
100
80
60
40
20
2 3 4
Figure 3 Frequency of postoperative atrial fibrillation (POAF)
based on quartiles of n-3 long-chain polyunsaturated fatty
acids (LC-PUFA) measured on the third postoperative day. See
Table 2 for the range of n-3 LC-PUFA levels (% of total fatty
acids) in each quartile.
6 ª 2011 The Authors. European Journal of Clinical Investigation ª 2011 Stichting European Society for Clinical Investigation Journal Foundation
G. V. SKULADOTTIR ET AL. www.ejci-online.com

7. from PL may be important for their biological role in inflam-
matory and electrophysiologic processes.
It has been proposed that the inflammatory response may be
a contributing factor in the development of POAF [3–6]. It has
therefore been speculated that short-term n-3 LC-PUFA treat-
ment may be effective in the prevention of POAF by shifting
the AA-derived eicosanoid pathway to the less-inflammatory
EPA-derived pathway [27]. Several studies have shown EPA-
and DHA-rich diets to be associated with low levels of pro-
inflammatory and high levels of anti-inflammatory markers
[28–30], while the lack of effect of dietary n-3 LC-PUFA on
inflammatory markers has also been reported [31]. The associa-
tion of POAF with the peak postoperative CRP concentration in
our study suggests that inflammation may have a role in the
(a)
(c) (d)
(f)
(g) (h)
(b)
(e)
Figure 4 Adjusted odds ratios and 95% confidence intervals of POAF for quartiles of preoperative and postoperative levels of fatty
acids in plasma phospholipids in patients undergoing coronary artery bypass graft surgery. (a) Preoperative and (b) postoperative
arachidonic acid (AA) levels; (c) preoperative and (d) postoperative eicosapentaenoic acid (EPA) levels; (e) preoperative and (f) post-
operative docosahexaenoic acid (DHA) levels; (g) preoperative and (h) postoperative n-3 long-chain polyunsaturated fatty acids
(LC-PUFA) levels. The models were adjusted for age, body mass index, smoking and maximal peak postoperative C-reactive
protein. See Table 2 for the range of fatty acid levels (% of total fatty acids) in each quartile.
European Journal of Clinical Investigation 7
PLASMA N-3 AND N-6 FATTY ACIDS AND POSTOPERATIVE ATRIAL FIBRILLATION

8. pathogenesis of this disorder [4,5,32]. However, no decrease in
the incidence of POAF in the patients undergoing CABG was
observed, despite relatively high n-3 LC-PUFA levels in their
plasma PL.
The relative balance between pro- and anti-inflammatory
factors may to some extent depend on baseline levels of AA and
EPA, nonesterified (free) in the circulation, or esterified in
plasma PL and in cell membrane PL, yielding a complex interac-
tion of these substances [25]. In addition, it has been suggested
that the ratio between the nonesterified AA and EPA in the cir-
culation and in cells could be more important than the AA and
EPA content or ratio in plasma PL and cell membranes for the
anti- and pro-inflammatory state in humans [8], as AA is metab-
olized to both pro- and anti-inflammatory eicosanoids [33].
The rate of POAF was relatively high in our study. This may,
at least in part, be related to our strict diagnostic criteria of only
a 5-min episode to define POAF as well as the use of continuous
electrocardiographic monitoring throughout their hospital stay
in all patients. However, other studies have also reported
similar rates of POAF [12].
Although the present study is prospective and well designed,
it is somewhat limited by the relatively small sample size. In a
larger sample, the association between plasma levels of n-3 LC-
PUFA and the risk of POAF could have been investigated more
thoroughly by comparing quintiles or deciles of n-3 LC-PUFA
levels. In particular, a larger study would be required to exam-
ine more closely the U-curve relationship between total n-3 LC-
PUFA in plasma PL and POAF suggested by our investigation.
Conclusions
The results of our study show that relatively high baseline or
postoperative levels of total n-3 LC-PUFA in plasma PL do not
prevent POAF following CABG surgery. A U-curve relation-
ship between total n-3 LC-PUFA in plasma PL and risk of
POAF may exist in the postoperative state and explain why
studies in populations with low baseline n-3 LC-PUFA levels
have demonstrated a beneficial effect. The inverse association
between AA levels and POAF is also of great interest. Whether
such an effect is mediated through anti-arrhythmic or anti-
inflammatory mechanisms remains unclear, although our study
supports a role for inflammation in the pathogenesis of POAF.
Determination of the role of dietary n-3 LC-PUFA in prevention
of POAF and in other arrhythmia management requires further
study. Better understanding of the pathogenesis of POAF is
essential for development of effective therapeutic strategies for
the prevention and treatment of this common disorder.
Acknowledgements
This work was supported by grants from the Icelandic
Centre for Research (RANNIS, Grant No. 080411021), the
University of Iceland Research Fund, and the Landspitali –
The National University Hospital of Iceland Research
Fund. The contribution of the participants, employees at
Landspitali – The National University Hospital of Iceland,
and Lilja G. Steinsdottir, Laboratory Assistant at the
University of Iceland, is gratefully acknowledged.
Contributions
All authors contributed to the design of the study. GVS, RH,
DOA and OSI collected and analysed the data; OSI performed
statistical analysis; GVS and OSI interpreted the data and wrote
the initial draft of the manuscript. All authors contributed to
the final form of the manuscript.
Conflict of interest
The authors declare no conflict of interest.
Address
Department of Physiology (G.V. Skuladottir, R. Heidarsdottir),
Faculty of Medicine, School of Health Sciences, University of
Iceland (G.V. Skuladottir, R. Heidarsdottir, D.O. Arnar, B.
Torfason, V. Edvardsson, R. Palsson); Division of Cardiology,
Department of Medicine, Cardiovascular Research Center
(D.O. Arnar, G. Gottskalksson); Department of Cardiothoracic
Surgery (B. Torfason); Children’s Medical Center
(V. Edvardsson); Division of Nephrology, Department of
Medicine (R. Palsson, O.S. Indridason), Landspitali – The
National University Hospital of Iceland, Hringbraut, Reykjavik,
Iceland.
Correspondence to: Gudrun V. Skuladottir, PhD, Department
of Physiology, Faculty of Medicine, School of Health Sciences,
University of Iceland, Vatnsmyrarvegur 16, IS-101
Reykjavik, Iceland. Tel: +354 525 4825; fax: +354 525 4886;
e-mail: gudrunvs@hi.is
Received 21 September 2010; accepted 28 January 2011
References
1 Kaireviciute D, Aidietis A, Lip GY. Atrial fibrillation following car-
diac surgery: clinical features and preventative strategies. Eur Heart
J 2009;30:410–25.
2 Maisel WH, Rawn JD, Stevenson WG. Atrial fibrillation after cardiac
surgery. Ann Intern Med 2001;135:1061–73.
3 Bruins P, te Velthuis H, Yazdanbakhsh AP, Jansen PG, van
Hardevelt FW, de Beaumont EM et al. Activation of the
complement system during and after cardiopulmonary bypass
surgery: postsurgery activation involves C-reactive protein and
is associated with postoperative arrhythmia. Circulation
1997;96:3542–8.
4 Chung MK, Martin DO, Sprecher D, Wazni O, Kanderian A, Carnes
CA et al. C-reactive protein elevation in patients with atrial arrhyth-
mias: inflammatory mechanisms and persistence of atrial fibrilla-
tion. Circulation 2001;104:2886–91.
8 ª 2011 The Authors. European Journal of Clinical Investigation ª 2011 Stichting European Society for Clinical Investigation Journal Foundation
G. V. SKULADOTTIR ET AL. www.ejci-online.com

9. 5 Aviles RJ, Martin DO, Apperson-Hansen C, Houghtaling PL, Rau-
taharju P, Kronmal RA et al. Inflammation as a risk factor for atrial
fibrillation. Circulation 2003;108:3006–10.
6 Lo B, Fijnheer R, Nierich AP, Bruins P, Kalkman CJ. C-reactive pro-
tein is a risk indicator for atrial fibrillation after myocardial revascu-
larization. Ann Thorac Surg 2005;79:1530–5.
7 Leaf A, Kang JX, Xiao Y-F, Billman GE. Clinical prevention of sud-
den cardiac death by n-3 polyunsaturated fatty acids and mecha-
nism of prevention of arrhythmias by n-3 fish oils. Circulation
2003;107:2646–52.
8 Calder PC. Polyunsaturated fatty acids and inflammatory processes:
new twists in an old tale. Biochimie 2009;91:791–5.
9 Calo` L, Bianconi L, Colivicchi F, Lamberti F, Loricchio ML, de Ruvo
E et al. N-3 Fatty acids for the prevention of atrial fibrillation after
coronary artery bypass surgery: a randomized, controlled trial. J Am
Coll Cardiol 2005;45:1723–8.
10 Heidt MC, Vician M, Stracke SK, Stadlbauer T, Grebe MT, Boening
A et al. Beneficial effects of intravenously administered N-3 fatty
acids for the prevention of atrial fibrillation after coronary artery
bypass surgery: a prospective randomized study. Thorac Cardiovasc
Surg 2009;57:276–80.
11 Heidarsdottir R, Arnar DO, Skuladottir GV, Torfason B, Edvardsson
V, Gottskalksson G et al. Does treatment with n-3 polyunsaturated
fatty acids prevent atrial fibrillation after open heart surgery?
Europace 2010;12:356–63.
12 Saravanan P, Bridgewater B, West AL, O’Neill SC, Calder PC,
Davidson NC. Omega-3 fatty acid supplementation does not reduce
risk of atrial fibrillation after coronary artery bypass surgery: a
randomized, double-blind, placebo-controlled clinical trial. Circ
Arrhythm Electrophysiol 2010;3:46–53.
13 Crowe FL, Allen NE, Appleby PN, Overvad K, Aardestrup IV, John-
sen NF et al. Fatty acid composition of plasma phospholipids and
risk of prostate cancer in a case-control analysis nested within the
European prospective investigation into cancer and nutrition. Am J
Clin Nutr 2008;88:1353–63.
14 Simera I, Moher D, Hoey J, Schulz KF, Altman DG. A catalogue of
reporting guidelines for health research. Eur J Clin Invest 2010;40:35–
53.
15 Folch J, Lees M, Sloane Stanley GH. A simple method for the isola-
tion and purification of total lipides from animal tissues. J Biol Chem
1957;226:497–509.
16 DeCaterina R, Giannessi D, Mazzone A, Bernini W, Lazzerini G,
Maffei S et al. Vascular prostacyclin is increased in patients ingesting
w-3 polyunsaturated fatty acids before coronary artery bypass graft
surgery. Circulation 1990;82:428–38.
17 Li Y, Kang JX, Leaf A. Differential effects of various eicosanoids on
the production or prevention of arrhythmias in cultured neonatal
rat cardiac myocytes. Prostaglandins 1997;54:511–30.
18 Abbatecola AM, Cherubini A, Guralnik JM, Andres Lacueva C,
Ruggiero C, Maggio M et al. Plasma polyunsaturated fatty acids and
age-related physical performance decline. Rejuvenation Res
2009;12:25–32.
19 Sadou H, Le´ger CL, Descomps B, Barjon JN, Monnier L, Crastes de
Paulet A. Differential incorporation of fish-oil eicosapentaenoate
and docosahexaenoate into lipids of lipoprotein fractions as related
to their glyceryl esterification: a short-term (postprandial) and long-
term study in healthy humans. Am J Clin Nutr 1995;62:1193–200.
20 Skuladottir GV, Gudmundsdottir S, Olafsson GB, Sigurdsson SB,
Sigfusson N, Axelsson J. Plasma fatty acids and lipids in two
separate, but genetically comparable, Icelandic populations. Lipids
1995;30:649–55.
21 Dyerberg J, Madsen P, Møller JM, Aardestrup I, Schmidt EB. Bio-
availability of marine n-3 fatty acid formulations. Prostaglandins
Leukot Essent Fatty Acids 2010;83:137–41.
22 Simopoulos AP. Evolutionary aspects of diet, the omega-6 ⁄ omega-3
ratio and genetic variation: nutritional implications for chronic
diseases. Biomed Pharmacother 2006;60:502–7.
23 Rapoport SI, Rao JS, Igarashi M. Brain metabolism of nutritionally
essential polyunsaturated fatty acids depends on both the diet and
the liver. Prostaglandins Leukot Essent Fatty Acids 2007;77:251–61.
24 Metcalf RG, James MJ, Gibson RA, Edwards JR, Stubberfield J,
Stuklis R et al. Effects of fish-oil supplementation on myocardial
fatty acids in humans. Am J Clin Nutr 2007;85:1222–8.
25 Simopoulos AP. The importance of the omega-6 ⁄ omega-3 fatty acid
ratio in cardiovascular disease and other chronic diseases. Exp Biol
Med 2008;233:674–88.
26 Nakamura H, Kim DK, Philbin DM, Peterson MB, Debros F, Koski
G et al. Heparin-enhanced plasma phospholipase A2 activity and
prostacyclin synthesis in patients undergoing cardiac surgery. J Clin
Invest 1995;95:1062–70.
27 Fritsche K. Fatty acids as modulators of the immune response. Annu
Rev Nutr 2006;26:45–73.
28 Ferrucci L, Cherubini A, Bandinelli S, Bartali B, Corsi A, Lauretani
F et al. Relationship of plasma polyunsaturated fatty acids to
circulating inflammatory markers. J Clin Endocrinol Metab
2006;91:439–46.
29 Micallef MA, Munro IA, Garg ML. An inverse relationship between
plasma n-3 fatty acids and C-reactive protein in healthy individuals.
Eur J Clin Nutr 2009;63:1154–6.
30 Farzaneh-Far R, Harris WS, Garg S, Na B, Whooley MA. Inverse
association of erythrocyte n-3 fatty acid levels with inflammatory
biomarkers in patients with stable coronary artery disease: The
Heart and Soul Study. Atherosclerosis 2009;205:538–43.
31 Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Willett WC,
Rimm EB. Habitual dietary intake of n-3 and n-6 fatty acids in
relation to inflammatory markers among US men and women.
Circulation 2003;108:155–60.
32 Tselentakis EV, Woodford E, Chandy J, Gaudette GR, Saltman AE.
Inflammation effects on the electrical properties of atrial tissue and
inducibility of postoperative atrial fibrillation. J Surg Res
2006;135:68–75.
33 Harris WS, Mozaffarian D, Rimm E, Kris-Etherton P, Rudel LL,
Appel LJ et al. Omega-6 fatty acids and risk for cardiovascular
disease: a science advisory from the American Heart Association
Nutrition Subcommittee of the Council on Nutrition, Physical
Activity, and Metabolism; Council on Cardiovascular Nursing;
and Council on Epidemiology and Prevention. Circulation
2009;119:902–7.
European Journal of Clinical Investigation 9
PLASMA N-3 AND N-6 FATTY ACIDS AND POSTOPERATIVE ATRIAL FIBRILLATION