Le´opold K. Fezeu1*, Franc¸ois Laporte2, Emmanuelle Kesse-Guyot1, Valentina A. Andreeva1, Jacques Blacher1,3, Serge Hercberg1,4, Pilar Galan1

1 Universite´ Paris 13, Sorbonne Paris Cite´ - Equipe de Recherche en Epide´miologie Nutritionnelle (EREN), Centre d’Epide´miologie et Biostatistiques (EPIBIOS), Inserm, Inra, Cnam, Universite´ Paris 5, Universite´ Paris 7, Bobigny, France, 2 De´partement de Biochimie–Pharmacologie–Toxicologie, Centre Hospitalier Universitaire, Grenoble, France, 3 Universite´ Paris-Descartes, Faculte´ de Me´decine, AP-HP; Hoˆtel-Dieu, Centre de Diagnostic et The´rapeutique, Paris, France, 4De´partement de Sante´ publique, Hoˆpital Avicenne, Bobigny, France

Introduction

The incidence of cardiovascular diseases (CVD) is suggested to be modulated by dietary fatty acid composition, ie saturated fatty acids (SFA) increasing and polyunsaturated fatty acids (PUFA) decreasing serum cholesterol levels and risk of CVD [1]. Some

fatty acids influence the incidence of CVD through triglyceride accumulation, inflammation, vasodilation, or platelet aggregation [2]. SFA and monounsaturated fatty acids (MUFA) are the main components of triglycerides, although mechanisms linking the relationships between CVD and individual fatty acids remain uncertain.

Table 1. Baseline sociodemographic, clinical and biologic characteristics of SU. FOL.OM3 participants without and with major events, cardiac and cerebrovascular events at the end of the follow-up.

Characteristics No events Major events P* Cardiac events P{ Cerebrovascular events P` N 1717 154 379 85 EPA + DPA supplementation, n (%)

1717 (50.9) 78 (50.6) 0.95 185 (48.8) 0.46 40 (47.1) 0.49

Age, years 61.468.9 63.669.5 0.006 62.369.3 0.12 66.068.5 ,0.0001 Men, n (%) 1375 (80.1%) 124 (80.5%) 0.67 315 (83.1%) 0.13 69 (81.2%) 0.43 BMI, kg/m2 27.564.0 27.864.4 0.27 27.764.0 0.48 28.064.1 0.19 SBP, mm Hg 133.1621.3 136.8623.4 0.24 134.1622.0 0.78 141.5623.9 0.03 DBP, mm Hg 83.4612.2 83.4612.5 0.87 82.5612.5 0.16 85.5612.2 0.20 MAP, mm Hg 99.9 14.5 101.2615.3 0.64 99.7614.8 0.51 104.2615.3 0.08 Total energy intake, kcal/day1

24456991 23436856 0.50 24706941 0.52 22996591 0.61

Total fibre intake, g/day1 24.6610.7 23.269.8 0.13 25.2610.6 0.42 24.069.3 0.52 Total fish intake, g/day1 62.3663.9 51.6639.6 0.07 63.7670.8 0.67 51.1644.4 0.20 Fat fish intake, g/day1 19.7626.9 17.3621.4 0.40 18.9622.0 0.65 17.0620.3 0.55 Total cholesterol, mmol/L 4.6061.05 4.7861.03 0.04 4.6861.09 0.24 4.8861.10 0.03 HDL-cholesterol, mmol/L 1.2060.30 1.2160.32 0.57 1.1960.32 0.22 1.2460.34 0.85 LDL-cholesterol, mmol/L 2.7660.84 2.8860.86 0.11 2.8460.90 0.19 2.9960.91 0.03 Triglycerides, mmol/L 1.4060.77 1.6061.00 0.004 1.5060.96 0.04 1.4660.88 0.36 Glycaemia, mmol/L 5.8161.56 6.1962.09 0.009 6.1262.09 0.001 5.9461.80 0.51 Smoking status, n (%) 0.007 0.23 0.002

Current smoker 181 (10.7) 25 (16.3) 36 (9.5) 14 (16.7) Former smoker 1040 (61.5) 92 (60.1) 249 (65.7) 52 (61.9) Never smoked 471 (27.8) 36 (23.5) 94 (24.8) 18 (21.4)

Alcohol intake, g/day 0.002 0.16 0.007

.10 833 (48.5) 71 (46.1) 181 (47.8) 40 (47.1) 0.5–10 468 (27.3) 27 (17.5) 88 (23.2) 13 (15.3) ,0.5 156 (9.1) 19 (12.3) 41 (10.8) 11 (12.9) Missing information 260 (15.1) 37 (24.0) 69 (18.2) 21 (24.7)

Inclusion criteria, n (%) 0.009 ,0.0001 ,0.0001

Stroke 444 (25.9) 59 (38.3) 55 (14.5) 52 (61.2) Myocardial infarction 773 (45.0) 63 (40.9) 198 (52.2) 18 (21.2) Unstable angina 500 (29.1) 32 (20.8) 126 (33.2) 15 (17.6)

Diabetes 0.004 0.002 0.03

Yes 278 (16.2%) 40 (26%) 87 (23.0%) 22 (25.9%) No 1439 (83.8%) 114 (74%) 292 (77.0%) 63 (74.1%)

Data are frequencies (percentages) or mean 6 SD. P values are for the comparison of *major, {cardiac and `cerebrovascular events with participants without events at the end of the study. 1Food frequency questionnaires were filled only by 85% of the study participants, and these values are obtained from subjects for which the information was available. doi:10.1371/journal.pone.0092548.t001

Based on these hypotheses, several epidemiological studies have examined associations between several fatty acids and stroke, arrhythmia and coronary disease risk factors. Some prospective epidemiologic studies have reported a significant association between fish intake or n–3 fatty acids derived from fish oil and a lower risk of CVD [1,3–6]. However, randomized controlled trials and meta-analyses have yielded inconsistent results [1,7–12]. A number of hypotheses could explain these discrepancies. First, the changes in tissue lipid composition as well as the beneficial antiinflammatory and anti-atherogenic properties of a diet high in n–3 fatty acids require a long time to express. Thus, studies and metaanalyses conducted for less than one year may be unable to detect

a clinical benefit [12]. Second, supplements given as part of dietary interventions have a strictly controlled dose and content unlike others; therefore, while compliance with a ‘‘self-administered’’ dietary supplement may be greater than that in dietary interventions, it is not clear whether both supplementations have similar effects. Many systematic reviews have included both dietary interventions and dietary supplements; this is likely to lead to significant heterogeneity [11]. Equally important is the fact that different formulations with different dosages have been used. As alinolenic acid is a plant-based n–3 polyunsaturated fatty acid (PUFA) that is poorly converted (less than 5%) to eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), it is therefore

1. Table 2. Baseline plasma lipid profile of SU.FOL.OM3 participants without and with major events, cardiac and cerebrovascular events at the end of the follow-up.

Plasma lipid profile No events Major events P* Cardiac events P{ Cerebrovascular events P` N 1717 154 379 85 Total fatty acids, mmol/L 1056862677 1130262920 0.003 10756.3 0.24 11056.1 0.19 Total saturated FA 29.7162.89 29.9963.26 0.52 29.7462.97 0.99 29.7363.19 0.77

Palmitic acid (C 16:0) 22.2862.28 22.5562.47 0.31 22.3362.37 0.85 22.3862.34 0.90 Trans fatty acids 0.1460.10 0.1460.10 0.98 0.1460.11 0.79 0.1460.10 0.91 Total MUFA 25.8363.73 26.5163.71 0.10 25.8963.79 0.90 26.2163.67 0.97

Oleic acid (C 18:1 n – 9) 21.3063.03 21.7563.16 0.24 21.3863.18 0.72 21.3963.11 0.51 Total PUFA 44.1365.66 43.1865.64 0.16 44.0565.65 0.96 43.7365.46 0.90 Total n-6 PUFA 38.7165.31 38.0765.31 0.44 38.7565.47 0.75 38.5965.01 0.59

Linoleic acid (C 18:2 n – 6) 27.2864.69 27.0764.57 0.95 27.5264.74 0.19 27.7064.65 0.16 Dihomo-U-linolenic acid

1.59 0.34 1.6260.40 0.30 1.5860.36 0.64 1.66 0.44 0.058

(C20:3 n – 6)

8.7961.96 8.3462.06 0.04 8.6062.15 0.06 8.2261.99 0.04

Arachidonic acid (C 20:4 n – 6)

Total n-3 PUFA 5.4261.84 5.1161.83 0.04 5.3061.89 0.29 5.1461.79 0.24 a-linolenic acid

1. (C 18:3 n – 3)
2. (C 18:4 n – 3)

0.0460.04 0.0360.03 0.006 0.0360.04 0.007 0.0260.03 0.005

EPA (C 20:5 n – 3) 1.4260.87 1.3360.85 0.16 1.3860.90 0.37 1.3360.88 0.42 DPA (C 22:5 n – 3) 0.6060.15 0.5760.15 0.03 0.5960.17 0.38 0.5860.14 0.10 DHA (C 22:6 n – 3) 2.7560.92 2.5860.93 0.04 2.6860.91 0.15 2.6360.88 0.33 EPA DHA 4.17 1.68 3.9161.67 0.06 4.06 1.70 0.21 3.96 1.66 0.34 EPA DPA + DHA 4.7761.76 4.4861.77 0.05 4.6561.81 0.21 4.5461.74 0.30

N – 3/n – 6 ratio 0.1460.05 0.1460.05 0.15 0.1460.06 0.37 0.1360.05 0.18

All values are expressed as percentage of total plasma fatty acids, except for total plasma fatty acids. P values are for the comparison of *major, {cardiac and `cerebrovascular events with participants without events at the end of the study, after adjusting for total energy intake. doi:10.1371/journal.pone.0092548.t002

unlikely that it would have the cardioprotective effects of fish oil [13,14].

Although several studies have examined the relationship between PUFA intake and the incidence of CVD, no clear consensus yet exists regarding the effects of various fatty acid compositions, including that of PUFA on the development of CVD [7,8,15]. This may be at least partly due to the fact that dietary intake is difficult to quantify precisely due to self-reporting bias, difficulties in evaluating consumed portion sizes, lack of information about cooking methods, and lack of precision as regards food composition tables, particularly for processed food [16,17]. Usual methods of dietary assessment such as food frequency questionnaires, diet history or repeated 24-h recall/ records are used in large epidemiological studies to assess fatty acids intake, yet these are subject to the above limitations, as well

1. as to errors in recall [16,17]. We recently completed the SU.FOL.OM3 study, a double

blind, randomized, placebo controlled secondary prevention trial which aimed to investigate whether dietary supplementation with B vitamins or n–3 fatty acids, or both, could prevent subsequent major cardiovascular events in patients with a history of CVD [18]. Supplementation with n-3 fatty acids did not have any significant effects on major vascular events. In the present ancillary analysis of the SU.FOL.OM3 trial, we explored the association between baseline plasma fatty acids (PFA) and the risk of

secondary cardiovascular events in this population with recent medical history of CVD.

Methods

The supporting CONSORT checklist is available as supporting information; see Checklist S1.

Ethics statement

The protocol was approved by the ethic commitee ‘‘Comite´ Consultatif pour la Protection des Personnes se preˆtant a` la Recherche Biome´dicale’’ (CCPPRB nu 1933) of Paris-Cochin Hospital, and the data protection board ‘‘Commission Nationale de l’Informatique et des Liberte´s’’ (CNIL nu 901230). Participants provided their written informed consent to participate in this study. The current analyses included 2,263 subjects for whom baseline PFA were available.

Study population

Details concerning study methods and participant characteristics were reported elsewhere [18]. In brief, 2,501 participants 45– 80 years (1,987 men and 514 women) with a prior history of an acute coronary or cerebral ischemic event occurring within 1 to 12 months were recruited via a network of 417 cardiologists, neurologists or other physicians throughout France. They were

1. Table 3. Multivariate hazard ratios (95% confidence intervals) showing the associations between baseline plasma lipid profile and incident major events in the SU.FOL.OM3 study.

Plasma lipid profile HR Q1 HR Q2 HR Q3 HR Q4 P P for trend Total fatty acids

1. Model 1 1.0 1.32 (0.80–2.17) 1.54 (0.95–2.50) 1.54 (1.08–2.81) 0.13 0.018
2. Model 2 1.0 1.18 (0.62–2.25) 1.37 (0.65–2.89) 1.06 (0.35–3.25) 0.68 0.15

Total SFA

1. Model 1 1.0 0.56 (0.35–0.91) 0.81 (0.52–1.25) 0.97 (0.64–1.48) 0.08 0.75
2. Model 2 1.0 0.58 (0.33–1.01) 0.82 (0.49–1.37) 0.71 (0.39–1.30) 0.27 0.42

Palmitic acid (C 16:0)

1. Model 1 1.0 0.58 (0.36–0.94) 0.88 (0.57–1.35) 0.94 (0.61–1.45) 0.14 0.80
2. Model 2 1.0 0.48 (0.27–0.86) 0.82 (0.49–1.37) 0.63 (0.34–1.17) 0.077 0.33

Trans fatty acids

1. Model 1 1.0 0.99 (0.64–1.54) 0.95 (0.61–1.50) 0.92 (0.58–1.45) 0.98 0.69
2. Model 2 1.0 0.95 (0.57–1.58) 1.03 (0.61–1.75) 0.90 (0.52–1.56) 0.96 0.80

Total MUFA

1. Model 1 1.0 1.74 (1.07–2.83) 1.52 (0.92–2.50) 1.88 (1.16–3.07) 0.066 0.029
2. Model 2 1.0 1.61 (0.91–2.84) 1.50 (0.83–2.71) 1.55 (0.79–3.04) 0.40 0.25

Oleic acid (C 18:1 n – 9)

1. Model 1 1.0 1.06 (0.66–1.70) 1.28 (0.82–2.01) 1.29 (0.82–2.03) 0.59 0.19
2. Model 2 1.0 1.12 (0.65–1.91) 1.04 (0.60–1.82) 0.94 (0.49–1.81) 0.94 0.834

Total PUFA

1. Model 1 1.0 0.79 (0.51–1.21) 0.83 (0.54–1.27) 0.64 (0.41–1.02) 0.31 0.089
2. Model 2 1.0 1.25 (0.71–2.21) 1.11 (0.59–2.09) 1.05 (0.53–2.08) 0.85 0.93

Total n-6 PUFA

1. Model 1 1.0 0.91 (0.59–1.41) 0.87 (0.56–1.34) 0.71 (0.45–1.13) 0.53 0.15
2. Model 2 1.0 0.95 (0.54–1.67) 0.99 (0.54–1.80) 0.93 (0.49–1.77) 0.99 0.87

Linoleic acid (C 18:2 n – 6)

1. Model 1 1.0 0.93 (0.58–1.48) 1.19 (0.77–1.85) 1.20 (0.77–1.88) 0.62 0.53
2. Model 2 1.0 0.96 (0.55–1.67) 0.99 (0.58–1.69) 1.34 (0.80–2.25) 0.55 0.28

Dihomo-U-linolenic acid (C 20:3 n – 6)

1. Model 1 1.0 0.96 (0.61–1.51) 0.99 (0.64–1.59) 1.02 (0.65–1.61) 0.99 0.88
2. Model 2 1.0 1.15 (0.66–2.02) 1.32 (0.77–2.27) 1.43 (0.82–2.49) 0.61 0.18

Arachidonic acid (C 20:4 n – 6)

1. Model 1 1.0 0.75 (0.49–1.12) 0.62 (0.40–0.96) 0.51 (0.32–0.81) 0.023 0.002
2. Model 2 1.0 0.91 (0.55–1.52) 0.84 (0.49–1.44) 0.60 (0.32–1.12) 0.424 0.133

Total n-3 PUFA

1. Model 1 1.0 0.89 (0.59–1.35) 0.69 (0.44–1.07) 0.65 (0.41–1.03) 0.184 0.032
2. Model 2 1.0 1.46 (0.89–2.39) 0.85 (0.48–1.52) 0.83 (0.46–1.50) 0.103 0.253

a-linolenic acid (C 18:3 n – 3)

1. Model 1 1.0 0.94 (0.61–1.45) 0.91 (0.59–1.43) 0.92 (0.59–1.44) 0.976 0.682
2. Model 2 1.0 0.96 (0.58–1.60) 0.88 (0.52–1.51) 0.89 (0.52–1.53) 0.989 0.612

Stearidonic acid (C 18:4 n – 3)

1. Model 1 1.0 1.14 (0.76–1.71) 0.80 (0.53–1.23) 0.66 (0.40–1.09) 0.152 0.062
2. Model 2 1.0 0.93 (0.58–1.51) 0.72 (0.44–1.18) 0.56 (0.30–1.01) 0.206 0.035

EPA (C 20:5 n – 3)

1. Model 1 1.0 0.95 (0.62–1.45) 0.83 (0.54–1.29) 0.78 (0.49–1.23) 0.692 0.233
2. Model 2 1.0 1.35 (0.81–2.24) 1.06 (0.61–1.84) 0.87 (0.48–1.59) 0.436 0.540

DPA (C 22:5 n – 3)

1. Model 1 1.0 0.78 (0.52–1.17) 0.58 (0.37–0.91) 0.64 (0.40–1.00) 0.076 0.019
2. Model 2 1.0 0.86 (0.52–1.43) 0.77 (0.45–1.30) 0.77 (0.44–1.34) 0.739 0.298

1. Table 3. Cont.

Plasma lipid profile HR Q1 HR Q2 HR Q3 HR Q4 P P for trend DHA (C 22:6 n – 3)

1. Model 1 1.0 1.13 (0.75–1.69) 0.77 (0.50–1.20) 0.52 (0.32–0.86) 0.014 0.004
2. Model 2 1.0 1.29 (0.78–2.13) 0.99 (0.57–1.71) 0.63 (0.33–1.19) 0.124 0.120

EPA + DHA

1. Model 1 1.0 0.91 (0.60–1.38) 0.73 (0.47–1.14) 0.62 (0.39–0.99) 0.182 0.028
2. Model 2 1.0 1.35 (0.82–2.22) 1.02 (0.58–1.80) 0.79 (0.43–1.45) 0.285 0.348

EPA + DPA + DHA

1. Model 1 1.0 0.96 (0.64–1.45) 0.77 (0.50–1.19) 0.59 (0.36–0.95) 0.120 0.019
2. Model 2 1.0 1.56 (0.95–2.58) 1.17 (0.67–2.04) 0.76 (0.40–1.43) 0.07 0.303

N – 3/n – 6 ratio

1. Model 1 1.0 0.88 (0.57–1.36) 0.96 (0.63–1.48) 0.72 (0.45–1.16) 0.559 0.258
2. Model 2 1.0 1.13 (0.68–1.86) 1.03 (0.61–1.74) 0.75 (0.42–1.33) 0.553 0.341

1. Model 1: adjusted for age, BMI, initial event, gender.
2. Model 2: adjusted for covariates in Model 1 + treatment allocation group (both placebos, B vitamin and placebo of omega 3, omega 3 and placebo of B vitamins, omega
3. 3 and B vitamins), tobacco smoking, alcohol consumption, systolic blood pressure, diabetes status, plasma cholesterol (HDL and LDL), triglycerides, total fatty acid intake, n-3 supplementation and homocysteinemia. doi:10.1371/journal.pone.0092548.t003

randomly assigned using a two-by-two factorial design, to receive either 1) B-vitamins alone, 2) n-3 PUFA alone 3) both active treatments or 4) both placebo treatments.

CVD ascertainment

Every six months, questionnaires were mailed to all participants seeking information about any serious disease outcomes, compliance with the assigned regimen, presence of cardiovascular risk factors and adverse effects of the supplementation. A nonmandatory visit was scheduled annually and participants who were unable or unwilling to attend or who did not return the questionnaire were interviewed by telephone by study physicians. Family physicians also contacted the trial participants every six months in order to seek information on any health events that may have occurred. Two independent committees of cardiologists or neurologists who were blinded to the treatment allocation adjudicated all events. Participants were censored either on the date of their first major event, the date of death, or, for those without an event, at the end of follow-up on July 1, 2009. Three composite endpoints were used to evaluate the association between baseline PFA and cardiovascular events: 1) Major cardiovascular events, which included non-fatal myocardial infarction, non-fatal ischemic stroke or death from cardiovascular disease. 2) Cardiac events, which included acute coronary syndrome without myocardial infarction; resuscitation from sudden death; coronary artery bypass surgery; coronary angioplasty; cardiac failure; ventricular arrhythmia; supraventricular arrhythmia; cardiac surgery of any kind and carotid surgery or caro0tid artery angioplasty. 3) Cerebrovascular events, which included: ischemic or haemorrhagic stroke and transient ischemic attack.

Assessment of PFA profile and data collected at baseline

Data on gender, age, smoking status, existing hypertension, diabetes or dyslipidaemia, n-3 supplementation and medication use were collected at baseline via a questionnaire. Dietary intake (especially fish consumption) and alcohol consumption were assessed using a Food Frequency Questionnaire. Body mass index (BMI in kg/m2) was calculated from the standardized measurements of height and weight. Blood pressure was recorded by

trained staff for each patient in the sitting position, after a 5-min rest.

Venous blood samples were drawn after at least a 5-hour fast. Measured biological parameters included cholesterol (total, LDL, and HDL) and triglycerides (measured by enzymatic method with the use of C8000 System Abbott, Rungis, France), fasting blood glucose concentrations (enzymatic method) and total homocysteine (measured by a competitive immunoassay with direct chemiluminescence detection). Biological assessments were obtained by standard methods at a centralized laboratory.

Total (free and esterified) fatty acids composition of plasma lipids was determined by gas chromatography of the corresponding methyl esters. Cleavage of the ester bonds of triglycerides and cholesterol esters was performed by saponification. Specifically, lipids were extracted from 150 mL aliquots of plasma with 8 mL hexane: isopropanol (3:2, vol:vol) after the addition of heptadecanoic acid as an internal standard. The fatty acids were saponified with 1.0 mL NaOH in dry methanol at 100uC for 10 min and methylated with boron trifluoride 14% in methanol. The methylated fatty acids were extracted with n-heptane and quantified by gas chromatography with a flame ionisation detector (6850 Agilent) using a capillary column (Quadrex 30 m length,

1. 0.25 mm id, film thickness 0.25 mm, Alltech). The temperature gradient was programmed at 50uC/min from 50uC to 140uC,
2. 1.4uC/min from 140uC to 165uC and 10uC/min from 165uC to 245uC/min. Hydrogen was the carrier gas. The identification of the individual methyl ester components was made by comparison with a mixture of commercial standards. The results were expressed as percentages of the total area of all fatty acid peaks.

Statistical analyses

Data are expressed as mean (6SD) or median (25th–75th percentiles) and percentages (95% CI) as appropriate. Comparison of baseline characteristics between participants with and without a recurrent cardiovascular event relied on Student t tests, Wilcoxon Rank tests or Chi square tests as appropriate. Cox proportional hazards models were used to estimate the hazard ratios (HRs) and 95% 95% CI for major cardiovascular, cardiac and cerebrovascular events by quartiles of each PFA as well as total saturated fatty

1. Table 4. Multivariate hazard ratios (95% confidence intervals) showing the associations between baseline plasma lipid profile and incident cardiac events in the SU.FOL.OM3 study.

Plasma lipid profile HR Q1 HR Q2 HR Q3 HR Q4 P P for trend Total fatty acids

Model 1 1.0 1.12 (0.84–1.51) 1.23 (0.93–1.65) 1.23 (0.92–1.65) 0.452 0.130 Model 2 1.0 0.97 (0.67–1.42) 1.17 (0.73–1.87) 1.00 (0.47–2.13) 0.670 0.635

Total SFA

Model 1 1.0 0.82 (0.62–1.10) 0.90 (0.68–1.19) 0.89 (0.67–1.18) 0.623 0.540 Model 2 1.0 0.76 (0.55–1.06) 0.90 (0.65–1.24) 0.77 (0.53–1.12) 0.339 0.301

Palmitic acid (C 16:0)

Model 1 1.0 0.78 (0.59–1.05) 0.85 (0.64–1.13) 0.85 (0.64–1.13) 0.401 0.362 Model 2 1.0 0.70 (0.50–0.97) 0.81 (0.59–1.12) 0.76 (0.52–1.10) 0.183 0.209

Trans fatty acids

Model 1 1.0 0.99 (0.75–1.31) 1.00 (0.75–1.34) 0.98 (0.73–1.31) 0.998 0.905 Model 2 1.0 0.97 (0.71–1.33) 1.02 (0.74–1.40) 0.94 (0.67–1.31) 0.966 0.789

Total MUFA

Model 1 1.0 1.03 (0.77–1.37) 1.04 (0.78–1.39) 1.06 (0.79–1.41) 0.987 0.718 Model 2 1.0 1.02 (0.74–1.41) 1.05 (0.75–1.50) 1.13 (0.77–1.67) 0.932 0.535

Oleic acid (C 18:1 n – 9)

Model 1 1.0 1.07 (0.80–1.42) 0.98 (0.73–1.31) 0.98 (0.73–1.31) 0.929 0.762 Model 2 1.0 1.06 (0.77–1.46) 0.97 (0.69–1.37) 1.00 (0.67–1.48) 0.959 0.873

Total PUFA

Model 1 1.0 1.02 (0.77–1.35) 1.01 (0.76–1.34) 1.02 (0.76–1.36) 0.999 0.935 Model 2 1.0 1.10 (0.78–1.56) 1.08 (0.74–1.57) 1.10 (0.73–1.66) 0.957 0.710

Total n-6 PUFA

Model 1 1.0 1.05 (0.79–1.39) 0.89 (0.66–1.20) 1.10 (0.83–1.46) 0.539 0.787 Model 2 1.0 1.10 (0.78–1.56) 0.93 (0.64–1.36) 1.16 (0.78–1.70) 0.594 0.650

Linoleic acid (C 18:2 n – 6)

Model 1 1.0 0.75 (0.56–1.00) 0.84 (0.64–1.11) 0.84 (0.63–1.12) 0.261 0.357 Model 2 1.0 0.85 (0.62–1.18) 0.92 (0.68–1.26) 0.86 (0.62–1.20) 0.754 0.443

Dihomo-U-linolenic acid (C 20:3 n – 6)

1. Model 1 1.0 0.88 (0.67–1.17) 0.74 (0.55–0.98) 0.93 (0.70–1.23) 0.209 0.342
2. Model 2 1.0 0.90 (0.68–1.19) 0.73 (0.54–0.98) 0.94 (0.71–1.29) 0.196 0.402

Arachidonic acid (C 20:4 n – 6)

1. Model 1 1.0 0.87 (0.66–1.14) 0.80 (0.60–1.07) 0.86 (0.65–1.14) 0.467 0.227
2. Model 2 1.0 0.89 (0.64–1.23) 0.90 (0.64–1.26) 0.89 (0.62–1.28) 0.882 0.564

Total n-3 PUFA

1. Model 1 1.0 0.99 (0.75–1.31) 0.82 (0.61–1.09) 0.89 (0.67–1.18) 0.480 0.241
2. Model 2 1.0 1.21 (0.88–1.67) 0.99 (0.71–1.40) 0.98 (0.70–1.38) 0.508 0.644

a-linolenic acid (C 18:3 n – 3)

1. Model 1 1.0 0.87 (0.65–1.17) 0.96 (0.72–1.28) 1.15 (0.87–1.51) 0.287 0.253
2. Model 2 1.0 0.90 (0.64–1.26) 1.04 (0.75–1.46) 1.28 (0.93–1.77) 0.172 0.079

Stearidonic acid (C 18:4 n – 3)

1. Model 1 1.0 1.10 (0.84–1.44) 0.83 (0.64–1.08) 0.67 (0.49–0.92) 0.013 0.005
2. Model 2 1.0 0.94 (0.69–1.29) 0.86 (0.65–1.15) 0.63 (0.44–0.90) 0.084 0.016

EPA (C 20:5 n – 3)

1. Model 1 1.0 0.91 (0.70–1.20) 0.68 (0.51–0.91) 0.79 (0.59–1.04) 0.054 0.027
2. Model 2 1.0 1.05 (0.77–1.43) 0.82 (0.59–1.15) 0.80 (0.56–1.13) 0.292 0.102

DPA (C 22:5 n – 3)

1. Model 1 1.0 0.88 (0.66–1.15) 0.82 (0.62–1.08) 0.79 (0.60–1.06) 0.365 0.091
2. Model 2 1.0 0.84 (0.61–1.15) 0.87 (0.63–1.19) 0.78 (0.56–1.09) 0.494 0.177

1. Table 4. Cont.

Plasma lipid profile HR Q1 HR Q2 HR Q3 HR Q4 P P for trend DHA (C 22:6 n – 3)

1. Model 1 1.0 1.11 (0.84–1.47) 1.03 (0.77–1.37) 0.74 (0.55–1.00) 0.0480 0.049
2. Model 2 1.0 1.34 (0.97–1.86) 1.25 (0.87–1.76) 0.84 (0.58–1.21) 0.0206 0.249

EPA + DHA

Model 1 1.0 0.90 (0.68–1.18) 0.81 (0.61–1.08) 0.82 (0.61–1.08) 0.426 0.122 Model 2 1.0 1.14 (0.82–1.57) 1.01 (0.72–1.42) 0.92 (0.65–1.30) 0.643 0.500

EPA + DPA + DHA

Model 1 1.0 0.95 (0.72–1.26) 0.82 (0.61–1.09) 0.83 (0.62–1.10) 0.409 0.117 Model 2 1.0 1.21 (0.88–1.67) 1.02 (0.71–1.41) 0.93 (0.66–1.33) 0.412 0.464

N – 3/n – 6 ratio

Model 1 1.0 0.71 (0.53–0.94) 0.73 (0.55–0.96) 0.80 (0.60–1.05) 0.0566 0.127 Model 2 1.0 0.81 (0.59–1.12) 0.85 (0.61–1.17) 0.85 (0.62–1.17) 0.582 0.377

1. Model 1: adjusted for age, BMI, initial event, gender.
2. Model 2: adjusted for covariates in Model 1 + treatment allocation group (both placebos, B vitamin and placebo of omega 3, omega 3 and placebo of B vitamins, omega
3. 3 and B vitamins),, tobacco smoking, alcohol consumption, systolic blood pressure, diabetes status, plasma cholesterol (HDL and LDL), triglycerides, total fatty acid intake, n-3 supplementation and homocysteinemia. doi:10.1371/journal.pone.0092548.t004

acids (SFA), trans fatty acids, monounsaturated fatty acids (MUFA), PUFA, n-6 PUFA, n-3 PUFA, EPA + DPA, EPA + DPA + DHA, and the n-3 PUFA to n-6 PUFA ratio. Associations with baseline plasma fatty acids were adjusted for two sets of potential confounding factors in model 1 (initial event, gender, baseline age and BMI) and model 2 (model 1 + treatment allocation group (both placebos, B vitamin and placebo of omega 3, omega 3 and placebo of B vitamins, omega 3 and B vitamins), tobacco smoking, alcohol consumption, diabetes status, systolic blood pressure, plasma cholesterol (total, HDL, LDL), n-3 supplementation, homocystinemia and total fatty acids intake). Statistical analyses were performed using STATA 10. Statistical tests were two-sided and p-values,0.05 were considered to be statistically significant.

Results

During the follow-up, among the 2,263 participants with available PFA, 154 developed a major cardiovascular event (median time to event, 2.3 years, 25th–75th percentiles: 1.3–3.3), 379 a cardiac event (1.3 years, 0.3–2.7) and 85 a cerebrovascular event (2.2 years, 1.2–3.2). The median follow-up of participants who did not develop a cardiovascular event until the end of the study was 4.8 (3.9–5.0) years. The major cardiovascular events included 57 stroke cases (among which 52 were non-fatal), 19 deaths from cardiac origin, 56 non-fatal myocardial infarctions and 14 sudden cardiac deaths. Cardiac events included 122 angioplasties, 51 myocardial infarctions, 45 cardiac arrhythmias, 40 coronary revascularisations and 40 cardiac interventions. Cerebrovascular events included 52 non-fatal ischemic stroke and 20 transient ischemic attacks.

Baseline characteristics

The mean baseline age of the participants was 61.6 (range: 41.9–81.2) years, and 80.6% of them were men. Regardless of the outcome considered, patients who developed a cardiovascular event more often had diabetes and were current smokers at baseline (except for cardiac events, Table 1). Participants with major cardiovascular events or with cardiac events had higher

plasma triglycerides, fasting plasma glucose and homocysteinaemia.

Baseline PFA and cardiovascular events

In bivariate analyses, stearidonic acid was lower in patients who developed a cardiovascular event, regardless the event considered. Total fatty acids were higher and arachidonic acid, total n–3 PUFA, DPA, DHA and the sum of EPA + DPA + DHA were lower in patients who developed a major cardiovascular event

1. (Table 2). Arachidonic acid was lower in patients who developed a cerebrovascular event.
2. (Table 3). This inverse association was borderline significant with increasing quartiles of stearidonic acid. The prevalence of major cardiovascular events was lower in the third and fourth quartiles of baseline plasma arachidonic acid (HR: 0.62, 0.40–0.96, p,0.03 and 0.51, 0.32–0.81, p,0.005 respectively) DPA (HR: 0.58, 0.37– 0.91, p,0.019 and 0.64, 0.40–1.00, p,0.05 respectively) and in the fourth quartiles of baseline plasma DHA (p,0.001), EPA + DHA (p,0.046) and EPA + DPA + DHA (p,0.029) respectively, in comparison with the first quartile. Furthermore, the risk of developing a cardiac event decreased with increasing quartiles of stearidonic acid, EPA and DHA (Table 4). The risk of developing a cerebrovascular event decreased with increasing quartiles of arachidonic acid and stearidonic acid (Table 5).

In model 2, only stearidonic acid remained inversely associated with the risk of developing a major cardiovascular event, a cardiac event or a cerebrovascular event, while arachidonic acid was inversely associated with the risk of cerebrovascular event.

Discussion

In this ancillary analysis which included patients with established and treated CVDs, plasma stearidonic acid and arachidonic acid showed protective effects on the recurrence of CVD and cerebrovascular diseases, respectively. The associations

1. Table 5. Multivariate hazard ratios (95% confidence intervals) showing the associations between baseline plasma lipid profile and incident cerebrovascular events in the SU.FOL.OM3 study.

Plasma lipid profile HR Q1 HR Q2 HR Q3 HR Q4 P P for trend Total fatty acids

Model 1 1.0 1.42 (0.73–2.79) 1.55 (0.80–3.00) 1.49 (0.77–2.91) 0.589 0.255 Model 2 1.0 0.47 (0.58–3.71) 1.61 (0.52–4.98) 1.27 (0.23–6.94) 0.691 0.621

Total SFA

Model 1 1.0 0.70 (0.39–1.26) 0.72 (0.40–1.30) 0.69 (0.38–1.26) 0.539 0.252 Model 2 1.0 0.71 (0.36–1.41) 0.85 (0.43–1.69) 0.53 (0.22–1.29) 0.496 0.246

Palmitic acid (C 16:0)

Model 1 1.0 0.86 (0.47–1.57) 0.85 (0.47–1.55) 0.83 (0.45–1.55) 0.932 0.578 Model 2 1.0 0.78 (0.38–1.61) 0.94 (0.46–1.90) 0.71 (0.29–1.72) 0.822 0.597

Trans fatty acids

Model 1 1.0 1.05 (0.59–1.87) 0.79 (0.41–1.49) 0.90 (0.49–1.65) 0.806 0.539 Model 2 1.0 0.93 (0.48–1.83) 0.67 (0.31–1.46) 1.02 (0.50–2.08) 0.721 0.820

Total MUFA

Model 1 1.0 1.84 (0.97–3.48) 1.24 (0.63–2.45) 1.93 (1.01–3.67) 0.127 0.141 Model 2 1.0 1.48 (0.71–3.05) 1.25 (0.58–2.67) 1.59 (0.67–3.82) 0.677 0.388

Oleic acid (C 18:1 n – 9)

Model 1 1.0 0.81 (0.43–1.51) 0.99 (0.55–1.77) 1.09 (0.61–1.95) 0.931 0.672 Model 2 1.0 0.78 (0.38–1.58) 0.74 (0.36–1.52) 0.79 (0.34–1.83) 0.847 0.505

Total PUFA

Model 1 1.0 0.81 (0.44–1.49) 1.02 (0.57–1.84) 0.85 (0.46–1.56) 0.840 0.782 Model 2 1.0 1.26 (0.56–2.85) 1.32 (0.55–3.18) 1.27 (0.50–3.27) 0.933 0.673

Total n-6 PUFA

Model 1 1.0 1.05 (0.57–1.93) 1.12 (0.62–2.03) 0.86 (0.46–1.62) 0.855 0.703 Model 2 1.0 1.00 (0.45–2.23) 1.22 (0.54–2.75) 1.05 (0.44–2.51) 0.948 0.826

Linoleic acid (C 18:2 n – 6)

Model 1 1.0 0.98 (0.54–1.77) 1.12 (0.62–2.01) 1.09 (0.59–2.01) 0.970 0.698 Model 2 1.0 0.93 (0.47–1.86) 0.64 (0.30–1.35) 1.03 (0.53–2.03) 0.624 0.789

Dihomo-U-linolenic acid (C 20:3 n – 6)

Model 1 1.0 0.96 (0.51–1.82) 1.11 (0.60–2.05) 1.35 (0.75–2.45) 0.675 0.275 Model 2 1.0 1.15 (0.51–2.59) 1.44 (0.66–3.13) 1.87 (0.88–3.98) 0.357 0.077

Arachidonic acid (C 20:4 n – 6)

Model 1 1.0 0.68 (0.39–1.16) 0.51 (0.28–0.92) 0.46 (0.24–0.86) 0.04 0.006 Model 2 1.0 0.69 (0.35–1.37) 0.52 (0.25–1.08) 0.44 (0.20–0.99) 0.18 0.033

Total n-3 PUFA

Model 1 1.0 1.04 (0.60–1.79) 0.69 (0.37–1.27) 0.69 (0.36–1.29) 0.37 0.125 Model 2 1.0 1.80 (0.94–3.46) 0.80 (0.36–1.77) 0.79 (0.35–1.77) 0.05 0.245

a-linolenic acid (C 18:3 n – 3) Model 1 1.0 1.09 (0.62–1.93) 0.98 (0.54–1.77) 0.90 (0.49–1.68) 0.95 0.7 Model 2 1.0 0.95 (0.48–1.88) 0.87 (0.43–1.77) 0.94 (0.46–1.94) 0.98 0.80

Stearidonic acid (C 18:4 n – 3)

Model 1 1.0 0.90 (0.54–1.49) 0.43 (0.23–0.80) 0.44 (0.22–0.89) 0.014 0.002 Model 2 1.0 0.76 (0.41–1.40) 0.38 (0.18–0.82) 0.50 (0.23–1.08) 0.058 0.014

EPA (C 20:5 n – 3)

Model 1 1.0 1.08 (0.61–1.94) 1.14 (0.64–2.03) 0.83 (0.44–1.57) 0.78 0.68 Model 2 1.0 1.40 (0.69–2.82) 1.28 (0.63–2.61) 0.84 (0.37–1.91) 0.53 0.72

DPA (C 22:5 n – 3)

Model 1 1.0 0.86 (0.49–1.51) 0.83 (0.47–1.47) 0.62 (0.33–1.17) 0.54 0.16 Model 2 1.0 0.91 (0.45–1.83) 1.05 (0.54–2.05) 0.59 (0.26–1.31) 0.50 0.31

Table 5. Cont.

Plasma lipid profile HR Q1 HR Q2 HR Q3 HR Q4 P P for trend DHA (C 22:6 n – 3)

Model 1 1.0 1.06 (0.61–1.84) 0.65 (0.36–1.19) 0.64 (0.34–1.20) 0.22 0.070 Model 2 1.0 1.04 (0.52–2.09) 0.90 (0.44–1.86) 0.65 (0.29–1.47) 0.68 0.29

EPA + DHA

Model 1 1.0 0.93 (0.52–1.64) 0.86 (0.48–1.53) 0.68 (0.36–1.27) 0.67 0.23 Model 2 1.0 1.21 (0.60–2.41) 1.12 (0.54–2.31) 0.73 (0.33–1.63) 0.62 0.47

EPA + DPA + DHA

Model 1 1.0 0.98 (0.56–1.71) 0.79 (0.44–1.42) 0.63 (0.33–1.19) 0.47 0.12 Model 2 1.0 1.48 (0.75–2.91) 1.06 (0.51–2.23) 0.69 (0.30–1.60) 0.27 0.31

N – 3/n – 6 ratio

1. Model 1 1.0 0.88 (0.48–1.59) 1.15 (0.66–2.00) 0.65 (0.34–1.26) 0.37 0.41
2. Model 2 1.0 1.13 (0.57–2.25) 1.22 (0.62–2.42) 0.71 (0.32–1.57) 0.56 0.54

1. Model 1: adjusted for age, BMI, initial event, gender.
2. Model 2: adjusted for covariates in Model 1 + treatment allocation group (both placebos, B vitamin and placebo of omega 3, omega 3 and placebo of B vitamins, omega
3. 3 and B vitamins),, tobacco smoking, alcohol consumption, systolic blood pressure, diabetes status, plasma cholesterol (HDL and LDL), triglycerides, total fatty acid intake, n-3 supplementation and homocysteinemia. doi:10.1371/journal.pone.0092548.t005

between total PFA, total n–3 PUFA, EPA, DPA, DHA, EPA + DHA, EPA + DPA + DHA and either major cardiovascular, cardiac or cerebrovascular events seemed to be mediated by established cardiovascular risks factors.

Numerous mechanistic studies have found that stearidonic acid favourably compares with dietary EPA in side-by-side experiments. Similar physiological effects were observed in ex vivo platelet aggregation studies, changes in tissue arachidonic acid content and eicosanoid formation, biomarkers of inflammation and modification of plasma lipid profiles. However, no epidemiological cohort study has tested the association between stearidonic acid with CVD. Moreover, stearidonic acid concentration is very low and at near the limit of detection, and dietary intakes is very low, thus the implications from a nutritional standpoint may be very limited. We also found evidence of an inverse association between arachidonic acid and CVD. Several studies [19–21] have shown an inverse association between plasma arachidonic acid and risk of CVD, while others have shown either no association [3,22]. Higher D5-desaturase index (a ratio of arachidonic to dihomo c-linolenic ratio) has been reported to be associated with higher insulin sensitivity [23,24], and lower incidence of the metabolic syndrome [25] or CVD [26], which may partly explain the inverse association between arachidonic acid and cardiovascular events observed here. However, these results contrast with other mechanistic studies revealing the prothrombotic and proinflammatory properties of arachidonic acid, and merit further exploration [27–30].

Our results suggest that associations with long chain n–3 PUFA are mainly mediated by confounding factors. Contrary to our findings, most [1,6,22,31,32], but not all [33–35] of the observational cohort studies show an inverse association between fish consumption, fish oil supplementation or plasma values of fish derived long chain n–3 PUFA and CVDs. A meta-analysis of such cohort studies [1] concluded that there was a modest evidence of an inverse association between n–3 PUFA intake and CVD, with no evidence of a linear dose-response relationship. Despite adjustment for potential confounders, residual confounding is likely to have occurred in some of these observational cohort studies. Indeed, participants who consumed more PUFA were

more likely to be non-smokers [33,36–39], to be more physically active [33,37,38], to take dietary vitamins [33], to eat more dietary fibre, fruit and vegetables, and less saturated and trans fats [33], and to drink less alcohol [33,37,38]. Not only did all these studies not take into account such potential confounding factors, but also, adjusting for these covariables does not fully protect against residual confounding. Whether fish consumption provides other beneficial nutrients not present in pure fish oil remains questionable. The most recently published meta-analyses of randomized controlled trials showed conflicting results regarding the effects of fish oil or fish supplements on the occurrence of primary CVD. Leon et al [8] reported on the effects of fish oil (EPA and DPA) on mortality and arrhythmias from 12 studies including 32,779 patients. No beneficial effect on arrhythmic events or all-cause mortality was reported but a significant reduction in deaths from cardiac causes was found. Marik et al [9] identified 11 studies that included 39,044 patients. They found that dietary supplementation with n–3 PUFA significantly reduced the risk of cardiovascular deaths, sudden cardiac deaths, and non-fatal cardiovascular events. With a total of 29 randomized controlled trials (n=35,144), with 25 of them concerning mortality and 14 restenosis, Filion et al [7] found that n–3 PUFA were not associated with mortality or with restenosis, even though the probability of some benefit remained high (0.93 and 0.90, respectively). Of the total population included in these metaanalyses, 20 to 35% were participants in the Gruppo Italiano per lo Studio della Sopravvivenza nell’Intarto Miocardio (GISSI)Prevenzione trial [40]. In this trial, the treatment under investigation was not blinded, and only a low or moderate proportion of patients received modern intervention and medical treatment. Also, only 5.0% of the patients had coronary revascularisation at baseline, and only 4.7% were on cholesterollowering drugs at hospital discharge. Therefore, it is not clear whether the conclusions from these meta-analyses are still valid in 2011. Moreover, three doubled-blind randomized controlled trials [15,18,41], not included in the previous meta-analyses, where almost 20,000 patients participated for a median follow up over 4 years, failed to demonstrate a protective effect of n–3 PUFA supplementation on the reduction of major cardiovascular events

in patients with personal history of CVD or prediabetes. The lack of association between baseline PFA and subsequent cardiovascular events in our study as well as in these two randomised controlled trials [15,18,41] could be due to an improvement in the efficacy of cardioprotective drug treatments between the 1995– 1996 and 2006–2007 [42]. There been improvement in survival but also the causes of death have shifted from CV to noncardiovascular origin [43]. Consequently, among patients who have had a CVD but who are receiving good clinical care and are

1. at relatively low risk of future CV events, a beneficial effect of habitual dietary intakes of fatty acids may be difficult to detect [44] as illustrated by findings from the Omega trial [15].

The principal strength of this study is its large, well-characterized population with close to complete follow-up with respect to clinical endpoints. Also, we adjusted for most traditional risks factors, therapeutic parameters, and biomarkers characterizing CVD. We evaluated not only the clinical effects of widely studied long chains PUFA (EPA and DHA), but also the effects of a vast majority of plasma fatty acids. Despite adjustment for potential confounders, residual confounding is likely to have occurred in that observational analysis. A major potential confounder is engagement in physical activity, which was not available for our participants at baseline. As subjects with high levels of physical activity eat significantly more fish compared with their more sedentary counterparts [33,37,38], it is likely that adjusting for levels of physical activity might have reduced the association between arachidonic acid or stearidonic acid and cardiovascular outcomes.

References

1. 1. Mente A, de Koning L, Shannon HS, Anand SS (2009) A systematic review of the evidence supporting a causal link between dietary factors and coronary heart disease. Arch Intern Med 169: 659–669.
2. 2. He K (2009) Fish, long-chain omega-3 polyunsaturated fatty acids and prevention of cardiovascular disease–eat fish or take fish oil supplement? Prog Cardiovasc Dis 52: 95–114.
3. 3. Harris WS, Poston WC, Haddock CK (2007) Tissue n-3 and n-6 fatty acids and risk for coronary heart disease events. Atherosclerosis 193: 1–10.
4. 4. Lee SH, Shin MJ, Kim JS, Ko YG, Kang SM, et al. (2009) Blood eicosapentaenoic acid and docosahexaenoic acid as predictors of all-cause mortality in patients with acute myocardial infarction–data from Infarction Prognosis Study (IPS) Registry. Circ J 73: 2250–2257.
5. 5. Lemaitre RN, King IB, Mozaffarian D, Kuller LH, Tracy RP, et al. (2003) n-3 Polyunsaturated fatty acids, fatal ischemic heart disease, and nonfatal myocardial infarction in older adults: the Cardiovascular Health Study. Am J Clin Nutr 77: 319–325.
6. 6. Sun Q, Ma J, Campos H, Rexrode KM, Albert CM, et al. (2008) Blood concentrations of individual long-chain n-3 fatty acids and risk of nonfatal myocardial infarction. Am J Clin Nutr 88: 216–223.
7. 7. Filion KB, El Khoury F, Bielinski M, Schiller I, Dendukuri N, et al. (2010) Omega-3 fatty acids in high-risk cardiovascular patients: a meta-analysis of randomized controlled trials. BMC Cardiovasc Disord 10: 24.
8. 8. Leon H, Shibata MC, Sivakumaran S, Dorgan M, Chatterley T, et al. (2008) Effect of fish oil on arrhythmias and mortality: systematic review. BMJ 337: a2931.
9. 9. Marik PE, Varon J (2009) Omega-3 dietary supplements and the risk of cardiovascular events: a systematic review. Clin Cardiol 32: 365–372.
10. 10. Saravanan P, Davidson NC, Schmidt EB, Calder PC (2010) Cardiovascular effects of marine omega-3 fatty acids. Lancet 376: 540–550.
11. 11. Hooper L, Thompson RL, Harrison RA, Summerbell CD, Moore H, et al.
12. 12. 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.
13. 13. Plourde M, Cunnane SC (2007) Extremely limited synthesis of long chain polyunsaturates in adults: implications for their dietary essentiality and use as supplements. Appl Physiol Nutr Metab 32: 619–634.
14. 14. Wang C, Harris WS, Chung M, Lichtenstein AH, Balk EM, et al. (2006) n-3 Fatty acids from fish or fish-oil supplements, but not alpha-linolenic acid, benefit cardiovascular disease outcomes in primary- and secondary-prevention studies: a systematic review. Am J Clin Nutr 84: 5–17.
15. 15. Kromhout D, Giltay EJ, Geleijnse JM (2010) n-3 fatty acids and cardiovascular events after myocardial infarction. N Engl J Med 363: 2015–2026.

In our population with established and well-treated history of ischemic heart disease or stroke, the inverse association of long chain n–3 PUFA with CVD was mainly driven by well-known cardiovascular risk factors. These results, as the lack of impact of our supplementation trial with DHA-EPA on CVD recurrence [18], do not support the recommendations of use of n-3 PUFA for the secondary prevention of CVD.

Supporting Information

Checklist S1 CONSORT Checklist. (DOC)

Acknowledgments

We thank the health professionals, staff at the coordinating centre and study centres as well as patients who participated in the SU.FOL.OM3 trial. We wish to thank Fabien Szabo de Edelenyi and Laurent Bourhis for their assistance with the database preparation.

Author Contributions

Conceived and designed the experiments: SH PG. Performed the experiments: SH PG. Analyzed the data: LKF. Contributed reagents/ materials/analysis tools: SH PG FL LKF. Wrote the paper: LKF. Assisted with the interpretation of data and read and edited each draft of the manuscript for important intellectual content: FL EKG VAA JB PG SH. Read and approved the final manuscript: FL EKG VAA JB PG SH.

1. 16. Arab L (2003) Biomarkers of fat and fatty acid intake. J Nutr 133 Suppl 3: 925S– 932S.
2. 17. Overby NC, Serra-Majem L, Andersen LF (2009) Dietary assessment methods on n-3 fatty acid intake: a systematic review. Br J Nutr 102 Suppl 1: S56–63.
3. 18. Galan P, Kesse-Guyot E, Czernichow S, Briancon S, Blacher J, et al. (2010) Effects of B vitamins and omega 3 fatty acids on cardiovascular diseases: a randomised placebo controlled trial. BMJ 341: c6273.
4. 19. Salonen JT, Salonen R, Penttila I, Herranen J, Jauhiainen M, et al. (1985) Serum fatty acids, apolipoproteins, selenium and vitamin antioxidants and the risk of death from coronary artery disease. Am J Cardiol 56: 226–231.
5. 20. Wang L, Folsom AR, Eckfeldt JH (2003) Plasma fatty acid composition and incidence of coronary heart disease in middle aged adults: the Atherosclerosis Risk in Communities (ARIC) Study. Nutr Metab Cardiovasc Dis 13: 256–266.
6. 21. Yamagishi K, Nettleton JA, Folsom AR (2008) Plasma fatty acid composition and incident heart failure in middle-aged adults: the Atherosclerosis Risk in Communities (ARIC) Study. Am Heart J 156: 965–974.
7. 22. Itakura H, Yokoyama M, Matsuzaki M, Saito Y, Origasa H, et al. (2011) Relationships between plasma fatty acid composition and coronary artery disease. J Atheroscler Thromb 18: 99–107.
8. 23. Borkman M, Storlien LH, Pan DA, Jenkins AB, Chisholm DJ, et al. (1993) The relation between insulin sensitivity and the fatty-acid composition of skeletalmuscle phospholipids. N Engl J Med 328: 238–244.
9. 24. Vessby B, Tengblad S, Lithell H (1994) Insulin sensitivity is related to the fatty acid composition of serum lipids and skeletal muscle phospholipids in 70-yearold men. Diabetologia 37: 1044–1050.
10. 25. Warensjo E, Riserus U, Vessby B (2005) Fatty acid composition of serum lipids predicts the development of the metabolic syndrome in men. Diabetologia 48: 1999–2005.
11. 26. Ohrvall M, Berglund L, Salminen I, Lithell H, Aro A, et al. (1996) The serum cholesterol ester fatty acid composition but not the serum concentration of alpha tocopherol predicts the development of myocardial infarction in 50-year-old men: 19 years follow-up. Atherosclerosis 127: 65–71.
12. 27. Bays HE, Tighe AP, Sadovsky R, Davidson MH (2008) Prescription omega-3 fatty acids and their lipid effects: physiologic mechanisms of action and clinical implications. Expert Rev Cardiovasc Ther 6: 391–409.
13. 28. Czernichow S, Thomas D, Bruckert E (2010) n-6 Fatty acids and cardiovascular health: a review of the evidence for dietary intake recommendations. Br J Nutr 104: 788–796.
14. 29. Harris W (2010) Omega-6 and omega-3 fatty acids: partners in prevention. Curr Opin Clin Nutr Metab Care 13: 125–129.
15. 30. Eussen SR, Geleijnse JM, Giltay EJ, Rompelberg CJ, Klungel OH, et al. (2012) Effects of n-3 fatty acids on major cardiovascular events in statin users and nonusers with a history of myocardial infarction. Eur Heart J.

1. 31. Aarsetoey H, Ponitz V, Grundt H, Staines H, Harris WS, et al. (2009) (n-3) Fatty acid content of red blood cells does not predict risk of future cardiovascular events following an acute coronary syndrome. J Nutr 139: 507–513.
2. 32. Mozaffarian D, Rimm EB (2006) Fish intake, contaminants, and human health: evaluating the risks and the benefits. JAMA 296: 1885–1899.
3. 33. 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.
4. 34. Manger MS, Strand E, Ebbing M, Seifert R, Refsum H, et al. (2010) Dietary intake of n-3 long-chain polyunsaturated fatty acids and coronary events in Norwegian patients with coronary artery disease. Am J Clin Nutr 92: 244–251.
5. 35. Oomen CM, Ocke MC, Feskens EJ, Kok FJ, Kromhout D (2001) alphaLinolenic acid intake is not beneficially associated with 10-y risk of coronary artery disease incidence: the Zutphen Elderly Study. Am J Clin Nutr 74: 457– 463.
6. 36. Egeland GM, Meyer HE, Selmer R, Tverdal A, Vollset SE (2001) Cod liver oil consumption, smoking, and coronary heart disease mortality: three counties, Norway. Int J Circumpolar Health 60: 143–149.
7. 37. Hu FB, Cho E, Rexrode KM, Albert CM, Manson JE (2003) Fish and longchain omega-3 fatty acid intake and risk of coronary heart disease and total mortality in diabetic women. Circulation 107: 1852–1857.
8. 38. Meyer KA, Kushi LH, Jacobs DR Jr, Folsom AR (2001) Dietary fat and incidence of type 2 diabetes in older Iowa women. Diabetes Care 24: 1528– 1535.

1. 39. Rissanen T, Voutilainen S, Nyyssonen K, Lakka TA, Salonen JT (2000) Fish oilderived fatty acids, docosahexaenoic acid and docosapentaenoic acid, and the risk of acute coronary events: the Kuopio ischaemic heart disease risk factor study. Circulation 102: 2677–2679.
2. 40. (1999) Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI-Prevenzione trial. Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto miocardico. Lancet 354: 447–455.
3. 41. Investigators OT, Bosch J, Gerstein HC, Dagenais GR, Diaz R, et al. (2012) n-3 fatty acids and cardiovascular outcomes in patients with dysglycemia. N Engl J Med 367: 309–318.
4. 42. Kotseva K, Wood D, De Backer G, De Bacquer D, Pyorala K, et al. (2009) Cardiovascular prevention guidelines in daily practice: a comparison of EUROASPIRE I, II, and III surveys in eight European countries. Lancet 373: 929–940.
5. 43. Roger VL, Weston SA, Gerber Y, Killian JM, Dunlay SM, et al. (2010) Trends in incidence, severity, and outcome of hospitalized myocardial infarction. Circulation 121: 863–869.
6. 44. Kromhout D, Yasuda S, Geleijnse JM, Shimokawa H (2012) Fish oil and omega3 fatty acids in cardiovascular disease: do they really work? Eur Heart J 33: 436– 443.