Red blood cell polyunsaturated fatty acids and mortality in the Women's Health Initiative Memory Study

Accepted Manuscript Red Blood Cell Polyunsaturated Fatty Acids and Mortality in the Women’s Health 1,2,3,4 Initiative Memory Study William S. Harris, PhD, Juhua Luo, PhD, James V. Pottala, PhD, Mark A. Espeland, PhD, Karen L. Margolis, MD. MPH, JoAnn E. Manson, MD, DrPH, Lu Wang, MD, PhD, Theodore M. Brasky, PhD, Jennifer G. Robinson, MD, MPH PII:

S1933-2874(17)30002-8

DOI:

10.1016/j.jacl.2016.12.013

Reference:

JACL 1039

To appear in:

Journal of Clinical Lipidology

Received Date: 23 September 2016 Revised Date:

15 December 2016

Accepted Date: 29 December 2016

Please cite this article as: Harris WS, Luo J, Pottala JV, Espeland MA, Margolis KL, Manson JE, Wang L, Brasky TM, Robinson JG, Red Blood Cell Polyunsaturated Fatty Acids and Mortality in the 1,2,3,4 Women’s Health Initiative Memory Study , Journal of Clinical Lipidology (2017), doi: 10.1016/ j.jacl.2016.12.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

1

Red Blood Cell Polyunsaturated Fatty Acids and Mortality in the Women’s Health Initiative

2

Memory Study1,2,3,4

3

William S. Harris, PhDa,b, Juhua Luo, PhDc, James V. Pottala, PhDa, Mark A. Espeland, PhDd,

5

Karen L. Margolis, MD. MPHe, JoAnn E. Manson, MD, DrPHf, Lu Wang, MD, PhDf, Theodore

6

M. Brasky, PhDg, and Jennifer G. Robinson, MD, MPHh

7

a. University of South Dakota, Sanford School of Medicine, Department of Internal Medicine,

10 11

b. OmegaQuant Analytics, LLC. Sioux Falls, SD

SC

9

Sioux Falls, SD

c. School of Public Health, Department of Epidemiology and Biostatistics, Indiana University, Bloomington, IN

M AN U

8

RI PT

4

12

d. Department of Biostatistical Sciences, Wake Forest School of Medicine, Winston-Salem, NC

13

e. HealthPartners Institute, Minneapolis, MN [email protected]

14

f. Brigham and Women’s Hospital, Harvard Medical School, Boston, MA

15

g. Comprehensive Cancer Center, Ohio State University, Columbus, OH

16

h. Department of Epidemiology, College of Public Health and Department of Internal Medicine, College of Medicine, University of Iowa, Iowa City, IA

18

TE D

17

Corresponding author

20

William S. Harris, PhD

21

OmegaQuant Analytics, LLC

22

5009 W. 12th St, Ste 8, Sioux Falls, SD 57106

23

605-271-6917

24

[email protected]

26 27 28

AC C

25

EP

19

Last names: Harris, Luo, Pottala, Espeland, Margolis, Manson, Wang, Brasky, Robinson

Running Title: RBC PUFAs and Mortality

1

ACCEPTED MANUSCRIPT

29

1

30

sponsor had no role in study design, study conduct, data analysis or manuscript preparation.

Financial Support: National Heart Lung and Blood Institute, Broad Area Announcement 19. The

31 32

2

33

the RBC fatty acid analyses were performed. None of the other authors have any potential

34

conflicts to disclose

RI PT

Conflict of Interest Disclosures: WSH is the President of OmegaQuant Analytics, LLC where

35 3

Online supporting materials found at (link)

38

4

Abbreviations: ALA, alpha-linolenic acid; ARA, arachidonic acid; CVD, cardiovascular

39

disease; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid;

40

FA, fatty acid; HR, hazard ratio; HT, hormone therapy; LA, linoleic acid; Omega-3 Index,

41

erythrocyte EPA+DHA; PUFA, polyunsaturated FA; RBC, red blood cell; SD, standard

42

deviation; WHIMS, Women’s Health Initiative Memory Study

36

M AN U

SC

37

AC C

EP

TE D

43

2

ACCEPTED MANUSCRIPT

ABSTRACT

45

Background: The prognostic value of circulating polyunsaturated fatty acid (PUFA) levels is

46

unclear.

47

Objectives: To determine the associations between red blood cell (RBC) PUFA levels and risk

48

for death.

49

Methods: This prospective cohort study included 6501 women aged 65-80 who participated in

50

the Women’s Health Initiative Memory Study (enrolment began 1996). RBC PUFA levels were

51

measured at baseline and expressed as a percent of total RBC PUFAs. PUFAs of primary interest

52

were the n-3 PUFAs eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) and their

53

sum (the Omega-3 Index). PUFA of secondary interest included the two major n-6 PUFAs,

54

linoleic (LA) and arachidonic (ARA), and the PUFA Factor score (a calculated variable

55

including 6 PUFAs that accounts for their inter-correlations). The primary outcome was total

56

mortality through August 2014.

57

Results: After a median of 14.9 years of follow-up, 1851 (28.5%) women had died. RBC levels

58

of EPA and DHA were higher in the survivors (p<0.002 for each). In the fully adjusted models,

59

the hazard ratios (99% confidence intervals) for mortality associated with a 1-SD PUFA increase

60

for total mortality were 0.92 (0.85, 0.98) for the Omega-3 Index, 0.89 (0.82, 0.96) for EPA, 0.93

61

(0.87, 1.0) for DHA, and 0.76 (0.64, 0.90) for the PUFA Factor score. There were no significant

62

associations of alpha-linolenic acid or ARA or LA with total mortality.

63

Conclusions: Higher RBC levels of marine n-3 PUFAs were associated with reduced risk for all-

64

cause mortality. These findings support the beneficial relationship between the Omega-3 Index

65

and health outcomes.

66

ClinicalTrials.gov Identifier: NCT00000611 (WHIMS).

67

Keywords: epidemiology; prospective cohort study; eicosapentaenoic acid; docosahexaenoic

68

acid; omega-3 fatty acids; omega-6 fatty acids.

AC C

EP

TE D

M AN U

SC

RI PT

44

69

3

ACCEPTED MANUSCRIPT

INTRODUCTION

71

Several recent studies have linked higher blood levels and/or dietary intakes of the long-chain n-

72

3 polyunsaturated fatty acids (PUFAs) with greater longevity. The fact that the Japanese have

73

both very high n-3 PUFA levels1, 2 and are among the longest-lived populations3 provides

74

ecological support to this hypothesis. Red blood cell (RBC) levels of eicosapentaenoic acid (EPA)

75

plus docosahexaenoic acid (DHA) are validated measure of n-3 PUFA status4, and this metric

76

(called the Omega-3 Index for simplicity), has shown association with death from cardiovascular

77

disease (CVD)5. In a previous investigation, the Omega-3 Index was inversely and independently

78

related to the rate of telomere attrition, a marker of cellular aging6. Among past studies that

79

examined the relations between PUFA biomarkers and risk for death4, 7-16 (Supplemental Table

80

1), most were conducted in Scandinavia, Europe or Taiwan, only three have been done in the US.

81

In addition, only two4, 12 utilized red blood cells (RBCs), and both were performed in patients

82

with known coronary heart disease (CHD). Hence the question remains open whether RBC

83

PUFA levels are associated with risk for all-cause and cause-specific death in the general US

84

population.

85

We recently analyzed the RBC PUFA composition in the Women’s Health Initiative Memory

86

Study (WHIMS)17-19. This is by far the largest cohort in which these exposure markers have been

87

measured with total mortality as an endpoint, hence we sought to explore the associations

88

between the Omega-3 Index with total mortality, and included secondary analyses for death from

89

CVD, cancer, and non-CVD, non-cancer (i.e., all other causes). As there have been concerns

90

regarding the role of the n-6 PUFAs in health20, we also examined the associations with the

91

major n-6 and n-3 FAs as well.

92

METHODS

93

Participants: The WHIMS randomized trial examined the effects of postmenopausal hormone

94

therapy (HT) on cognitive function in women aged 65-80 years21. Recruitment began in 1996.

95

There were 7299 enrolled in WHIMS with RBC FA measurements at baseline. The following

96

were excluded: 17 were lost to follow-up, 213 had RBC FA data that was technically unusable22,

97

and 568 were missing relevant covariates. A total of 6501 women, of whom 1851 (28.5%) died

98

during follow up, were included in this analysis. All WHIMS participants gave informed consent,

99

and Institutional Review Boards at all participating centers approved the study protocol.

AC C

EP

TE D

M AN U

SC

RI PT

70

4

ACCEPTED MANUSCRIPT

Covariates and Mortality Outcomes: Participants self-reported demographic, health history,

101

and lifestyle factors on standardized questionnaires at baseline. Variables and their definitions

102

are given in Table 1. Aspirin use and medical histories were collected by self-report. Mortality

103

outcomes included deaths from all causes; and those due to CVD, cancer, and all other causes.

104

The WHI data were linked with the National Death Index of the National Center for Health

105

Statistics. Causes of death were centrally adjudicated by trained physicians as described23. All

106

women were followed from their date of enrollment to their date of death, or August 29, 2014,

107

whichever came first.

108

RBC Fatty Acid Analysis: RBC membrane FA composition was analyzed using gas

109

chromatography and was expressed as a weight percent of total identified FAs 22. The inter-

110

assay coefficient of variation for the primary PUFAs of interest here (see below) was <6.5%.

111

During the aliquoting phase, the RBC samples were stored at -20°C for a period of

112

approximately two weeks, causing oxidative degradation of the PUFAs before measurement. The

113

original FA levels were estimated with multiple imputations as described. 22

M AN U

SC

RI PT

100

114

Statistical analysis: The primary PUFAs of interest in this study were EPA, DHA, and the

116

Omega-3 index. PUFA of secondary interest included alpha-linolenic acid (ALA),

117

docosapentaenoic acid n-3 (DPAn-3), linoleic acid (LA), arachidonic acid (ARA), and two

118

simple ratios (n-6/n-3 and ARA/EPA), and the PUFA factor score. The latter (specific to women)

119

included a weighted combination of six RBC PUFAs: RBC PUFAs: 0.37*lnALA + 0.84*lnEPA

120

+ 0.80*DHA - 0.667*ARA - 0.814*adrenic - 0.788*DPAn-6)24. Accordingly, the PUFA factor

121

score is more complex (because it accounts for inter-correlations among the PUFAs) but

122

directionally inverse compared with the two ratios. Continuous variables were summarized using

123

the mean ± SD and categorical variables were summarized using frequency (%). Student’s t-test,

124

one-way ANOVA, and the Wilcoxon Rank sum test were used to compare continuous variables

125

across groups. The chi square test was used to compare categorical variables across groups.

126

Unadjusted cumulative hazard estimates of total mortality by PUFA quartiles were calculated by

127

the Nelson-Aalen method. Cox proportional hazards regression models were used to calculate

128

the association HR (99% CI) between RBC PUFAs and all-cause mortality (primary outcome),

129

cancer mortality, CVD mortality and all other causes of death (secondary outcomes). HT

AC C

EP

TE D

115

5

ACCEPTED MANUSCRIPT

treatment assignments (estrogen/placebo, or estrogen + progesterone/placebo) in the two

131

WHIMS trials were included as four strata in the Cox regression models. The natural logarithms

132

of the PUFAs were used to reduce skew (the PUFA Factor score was normally distributed), and

133

then standardized to report the risk associated with a 1 standard deviation (SD) increase in each

134

biomarker. Model 1 was adjusted for age and race. Model 2 was further adjusted for BMI,

135

educational attainment, smoking pack-year, physical activity, weekly alcohol intake, waist

136

circumference, region, family history of cancer, family history of CVD, and aspirin use. Model 3

137

was further adjusted for high cholesterol requiring pills (ever), hypertension, and diabetes, and is

138

considered the fully-adjusted model. The sensitivity analyses included the exclusion of deaths

139

that occurred in the first 2 years of follow-up (n=65), the exclusion of participants with prevalent

140

CVD or cancer at baseline (n=1312), the stratification by use of aspirin and of cholesterol-

141

lowering pills, and the additional inclusion of the normalized socioeconomic status in the fully

142

adjusted models. In addition, the control and active hormone therapy groups were analyzed

143

separately for associations between the Omega-3 Index and total mortality. A separate analysis

144

of the Omega-3 Index and total mortality was conducted based on proposed clinical cut-points

145

for this metric (<4%, 4%-8%, and >8%)5 in the fully adjusted model. To control somewhat for

146

multiple testing, a critical level of 0.02 was used to determine a statistically significant

147

association between the Omega-3 Index, RBC EPA, DHA and all-cause mortality. A critical

148

value of 0.01 was used for all secondary analyses. Both critical values were chosen a priori to

149

account for multiple testing. SAS (Cary, NC) version 9.4 was used for all analyses.

150

RESULTS

151

Cohort description

152

A comparison of the characteristics of the women who remained alive or died during the 14.9-

153

year (median, 12.9-15.9) follow-up period are given in Supplemental Table 2. The women who

154

died were about 2 years older at baseline, they were less active and smoked more, and they were

155

more likely to have major risk factors for CVD (hypertension, diabetes). The most common sites

156

for fatal cancers were lung (32%), colon (10%), multiple myeloma and breast (5.4% each). The

157

most common causes of CVD mortality were definite or possible CHD (48%), and

158

cerebrovascular disease (25%); and for other deaths, chronic obstructive pulmonary disease

159

(11%), Alzheimer’s disease (10%), and other dementia (9%).

AC C

EP

TE D

M AN U

SC

RI PT

130

6

ACCEPTED MANUSCRIPT

Participant characteristics by Omega-3 Index quartile are given in Table 1. A higher Omega-3

161

Index was associated with older age, having a uterus (i.e., being assigned to the

162

estrogen+progesterone, as opposed to the estrogen-alone, arms), greater alcohol intake, higher

163

education, more physical activity, more frequent use of cholesterol-lowering medications, less

164

smoking, lower waist/BMI, and being non-Hispanic white. The Omega-3 Index was highest in

165

the Northeast and lowest in the Midwest.

166

RBC PUFA and Total Mortality

167

The median levels of the Omega-3 Index (and its constituents and DPAn-3) were higher in the

168

women who survived versus those who died (Table 2). There were no group-wise differences in

169

median levels of LA, ALA, or ARA, but all metrics that included both n-3 and n-6 PUFAs

170

differed significantly.

M AN U

SC

RI PT

160

171

For total mortality, a 1-SD greater Omega-3 Index was associated with an 8% risk reduction in

173

the fully-adjusted model (Table 3). Risk reductions of 7%-11% were seen for the three

174

individual long-chain n-3 PUFAs (EPA, DHA, and DPAn-3). Similar findings held for all

175

metrics that included these PUFAs, i.e., the PUFA Factor score, ARA/EPA and n-6/n-3 ratios.

176

Neither of the n-6 PUFAs (LA or ARA) was related to mortality.

TE D

172

177

Cumulative Nelson-Aalen unadjusted hazard estimates of total mortality by quartiles of the

179

Omega-3 Index and the PUFA factor score are shown in Figure 1. Fully adjusted hazard ratios

180

by quartile are shown in Figure 2 for the Omega-3 Index, EPA, DHA, LA and the PUFA factor.

181

Significant (<0.01) trends were observed for all but LA. Risk for death from any cause was

182

significantly lower at an Omega-3 Index >8% vs <4% (HR, 0.69; 95% CI 0.52 to 0.93, p=0.017),

183

and the trend across the 3 categories (<4%, 4-8% and >8%) was significant (p=0.019)

184

(Supplemental Figure 1).

AC C

185

EP

178

186

7

ACCEPTED MANUSCRIPT

In sensitivity analyses (Supplemental Table 3) none of the associations observed in the fully

188

adjusted model were altered by excluding deaths in the first two years, or after excluding

189

participants with a history of CVD or cancer [except the association with DHA was attenuated

190

(p>0.01)]. A stratified analysis with those women not taking aspirin revealed statistically

191

significant a protective effect between Omega-3 Index, EPA, the PUFA factor score, the two

192

ratios and all-cause mortality in the fully adjusted model, but no association was seen with the

193

smaller number of women who were taking aspirin. An additional sensitivity analysis was done

194

for use of cholesterol-lowering drugs. The HRs (99% CI) for the 5344 not taking these drugs was

195

0.92 (0.85, 1) and 0.90 (0.76, 1.07) for the 1157 who were. The association between the Omega-

196

3 Index and all-cause mortality in each HT treatment arm was examined. HRs (99% CIs) were

197

0.91 (0.82, 1.0) in the control arm and 0.93 (0.84, 1.0) in the active arm (P for interaction =.93).

198

Finally, the addition of socioeconomic status to the fully adjusted model for all-cause mortality

199

did not materially affect the HR (0.92 to 0.92), nor did stratification by college degree or race

200

(non-Hispanic white or not).

201 202

RBC Polyunsaturated Fatty Acids and Cause-Specific Mortality

203

Associations between PUFAs and CVD mortality, cancer mortality, and death from all other

204

causes were examined only in the fully-adjusted models. Only EPA was significantly (p<0.01)

205

associated with CVD mortality with an HR (99% CIs) of 0.88 (0.77, 1.00) (Supplemental Table

206

4). The lowest HR was seen with the PUFA Factor score which was marginally significant [0.76

207

(0.57, 1.01), p=0.012)].

TE D

EP

208

M AN U

SC

RI PT

187

DISCUSSION

210

In this study of older post-menopausal women, the baseline Omega-3 Index was significantly

211

lower in those who died during follow-up than those who did not. The same is true for the two

212

components of the Index, EPA and DHA, as well as for DPAn-3 and for all metrics that included

213

n-3 PUFA levels. Since neither of the n-6 PUFAs examined here (LA or ARA) were related with

214

vital status, the differences in summary ratios were primarily influenced by the n-3 PUFA

215

components. These findings were robust in four sensitivity analyses. In secondary analyses by

216

cause of death, EPA was inversely associated with risk for CVD death, but DHA was not, and a

217

higher PUFA factor score correlated with lower risk for non-CVD, non-cancer deaths.

AC C

209

8

ACCEPTED MANUSCRIPT

Several previous studies linking n-3 PUFA with total mortality (including the present one) are

219

summarized are in Supplemental Table 1 where they are categorized as epidemiologic (using

220

diet, supplement use, and circulating biomarker as exposure indicators), and as randomized trials.

221

In general, n-3 PUFA exposures based on dietary records (citations in the Table) have been

222

inconsistently associated with reduced mortality, however all three studies based on reported fish

223

oil use reported lower death rates with n-3 PUFA supplementation. In nine biomarker-based

224

studies, total mortality was inversely associated with n-3 PUFA levels in eight of the nine studies

225

that addressed the question. The only exception was a small, case-control study where

226

cholesteryl ester FA composition was used as the biomarker of exposure8. This study may have

227

failed to find an association because, of all the lipid classes, this is the only one where DHA is

228

present at lower levels than EPA, and where fish oil feeding has the least effect on DHA levels25.

229

Among the other n-3 PUFAs of interest, neither ALA nor DPAn-3 were significantly related to

230

overall mortality, although the HR for the latter (0.94, p<0.014) was very close to the critical

231

value of <0.01. Previous studies have also observed inverse associations between risk for CVD

232

death and DPAn-3 and ALA26 . In light of this and other prior reports (e.g. in heart failure27), our

233

observations can be considered supportive of a beneficial association between DPAn-3 and total

234

mortality. Our null findings for ALA agree with some previous studies8, 14, but ALA levels have

235

been associated with CVD death26, and ALA intake was associated with reduced total mortality

236

in a large RCT from Spain28.

237

TE D

M AN U

SC

RI PT

218

N-6 PUFA

239

As regards the other PUFAs of interest in this study, LA was not related to any mortality

240

outcome. This contrasts with the findings of three other studies7-9 that showed inverse

241

associations with total mortality using plasma cholesteryl ester, phospholipid or total lipid LA

242

levels as exposure markers. The different PUFA pools used in these studies may be part of the

243

explanation for this inconsistency. The correlation between RBC LA and plasma phospholipid

244

LA is 0.71, and cholesteryl ester LA is 0.65 (compared with 0.93 and 0.83, respectively for

245

EPA+DHA; unpublished data) might explain why our data agree with theirs for the n-3 PUFAs

246

but not for LA. In a prior study involving many sample types, LA levels were related to non-fatal,

247

but not fatal CVD events29. In addition, Wu et al. found plasma phospholipid levels of LA (but

248

not ARA) were related to total and CVD mortality in the Cardiovascular Health Study7. The

AC C

EP

238

9

ACCEPTED MANUSCRIPT

249

ARA (and long chain n-3 PUFA) findings comport with ours, but their LA data do not. The basis

250

for these inconsistent findings is unknown.

251

Combined PUFA metrics

253

Of the composite PUFA metrics, the PUFA factor score, a complex calculation combining both

254

n-3 and n-6 PUFAs, was associated more strongly with total mortality than any of the other RBC

255

PUFA metrics examined here. This metric was derived by Pottala et al. via structural equation

256

modeling using RBC PUFA data from the Framingham Offspring cohort24. These findings imply

257

that care must be taken to include the right FAs in a combined metric. For example, LA – which

258

is a major driver of the n-6:n-3 ratio – is not in the PUFA Factor score, and this may have

259

contributed to the superior predictive power of this metric. The reason for this may be that LA is

260

actually protective (based on the findings of 4 previous studies7-9, 14; it was neutral in the present

261

study), not harmful; and to include a “protective” FA in with other “harmful” FAs in the same

262

term (total n-6) adds unnecessary noise and obscures the signal. The fact that the simpler

263

ARA/EPA metric had similar predictive power as the n6:n3 ratio despite its non-inclusion of LA,

264

suggests that the other FAs in the Factor score that are not present in this simple ratio (e.g., DHA,

265

and the longer-chain products of ARA) appear to contribute to its improved predictive value. The

266

other difference between the PUFA Factor score and these simpler ratios is the weights given to

267

the FAs. In the former, specific weight-based on intercorrelations (that were sex-specific) were

268

derived experimentally, whereas in the latter ratios, all FAs are essentially given the same weight.

269

These are but two of the many problems with metrics like the n-6:n-3 ratio30. Further studies

270

with the PUFA Factor score are clearly needed.

SC

M AN U

TE D

EP

AC C

271

RI PT

252

272

Cause specific mortality

273

Many studies have documented reduced CVD risk associated with higher n-3 PUFA intakes

274

and/or circulating levels31. In the present study, we found non-significant associations (p>0.01)

275

with the Omega-3 Index and its major component, DHA, but a significant 12% reduction in

276

CVD mortality per SD increase in EPA levels, and a 25% reduction for the PUFA factor score.

277

This finding is consistent with a previous study in post-MI patients where RBC EPA, but not

278

DHA, levels significantly improved prediction of 2-year mortality over a composite risk

279

algorithm12. On the other hand, both EPA and DHA were associated with lower CVD mortality 10

ACCEPTED MANUSCRIPT

in another study of elderly individuals in the US10. This study measured plasma phospholipid n-3

281

PUFAs and included elderly men and women. It seems unlikely that the differences in

282

participant characteristics and PUFA biomarker measurements can explain the divergent results.

283

We observed no significant associations between RBC n-3 PUFA and cancer mortality. Few

284

others have examined associations between long-chain n-3 PUFA or fish intake and cancer

285

mortality32-34. In the only prior study using long-chain n-3 PUFA biomarkers, there was a

286

significant inverse relation between plasma phospholipid DPA and cancer mortality, but none for

287

EPA or DHA10. Three prospective cohorts examined dietary plus supplements of n-3 PUFA34 or

288

fish intake and cancer mortality32, 33. n-3 PUFA intake was inversely associated with cancer

289

death, whereas no association was seen with fish intake. Here, the HRs for deaths from causes

290

other than CVD and cancer were marginally significantly inversely associated with the PUFA

291

factor score, the Omega-3 Index and its component PUFAs.

M AN U

SC

RI PT

280

292

Mechanisms that explain the associations between higher RBC n-3 PUFA and improved

294

longevity are not clearly understood, but there are beneficial effects of n-3 PUFA on a variety of

295

CV risk factors that may be involved. These include reductions in serum triglyceride levels35,

296

blood pressure36, platelet aggregation37, heart rate38, susceptibility to ventricular fibrillation (in

297

some settings)39, inflammatory markers40, 41, plaque vulnerability42, 43 along with improvements

298

in endothelial function44. The observation that whole blood EPA+DHA levels are independently

299

associated with slower rates of telomere shortening6, a purported marker of “cellular aging,” is

300

yet another favorable biomarker relationship, but it is not a molecular mechanism.

EP

301

TE D

293

It is of interest to consider how much EPA+DHA one might have to consume to increase the

303

Omega-3 Index from Q1 (median 3.6%) to Q4 (7.1%). Based on a recent dose – response study45,

304

such an increase would be expected from an increase of about 1 g of EPA+DHA per day. To

305

achieve this daily intake of EPA+DHA (which is similar to the mean intake in Japan in 199046)

306

would require the daily consumption of about 2-3 oz. of salmon or about a pound of lower-fat

307

seafood like chunk light tuna, shrimp or flounder. Alternatively, 3 standard (300 mg EPA+DHA)

308

fish oil pills per day would suffice. Several early omega-3 RCTs used roughly this dose and

309

reported reduced total mortality47-49, whereas other more recent studies at the same dose have not

310

confirmed this finding50-52.

AC C

302

11

ACCEPTED MANUSCRIPT

311

The question of whether increased intakes of EPA+DHA will prolong life remains unsettled.

313

Clearly observational studies alone cannot answer the question of whether higher (vs lower) n-3

314

PUFA tissue levels (as reflected by the Omega-3 Index53-55) will, over the long-term, lower risk

315

for death. Though early randomized controlled trials with n-3 PUFAs found reduced overall

316

mortality47, 48, other more recent studies51, 52, 56, 57 have not. Some of the potential reasons why

317

these studies may have failed (e.g., better acute medical interventions, pervasive use of statins58,

318

short follow-up periods, low doses, etc.), have been reviewed59, 60. At present, there are four

319

major n-3 PUFA RCTs ongoing: The Reduction of Cardiovascular Events with EPA ‒

320

Intervention Trial (REDUCE-IT, Clinicaltrials.com code NCT01492361), Statin Residual Risk

321

Reduction with Epanova in High Cardiovascular Risk Patients with Hypertriglyceridemia

322

(STRENGTH, NCT02104817), the Vitamin D and Omega-3 Trial (VITAL, NCT0116925961),

323

and A Study of Cardiovascular Endpoints in Diabetes (ASCEND, NCT00135226). The details of

324

each trial have been recently summarized62. Of these event-driven trials, REDUCE-IT and

325

STRENGTH are testing pharmacological dose (2-4 g/day) of EPA (4 g/day) or EPA+DHA

326

carboxylic acids, respectively, for 4-6 years in patients with hypertriglyceridemia and on statin.

327

ASCEND and VITAL are both primary prevention, 2x2 factorial trials testing, in the former, the

328

effects of 700 mg of EPA+DHA per day (with or without aspirin) on CVD endpoints in about

329

15,000 patients with diabetes; and in the latter, the effects of 840 mg of EPA+DHA per day (with

330

or without vitamin D) on CVD and cancer endpoints in about 25,000 adults. Results of these

331

studies will further our understanding of the role that n-3 PUFA may play in health outcomes.

SC

M AN U

TE D

EP

332

RI PT

312

Strengths and Limitations

334

Strengths of this study include the large sample size and number of events, the unambiguous

335

nature of the primary endpoint, and the use of an objective biomarker of PUFA exposure with

336

low biological variability63. Limitations would include the unique nature of the study population

337

- postmenopausal women, average age 70, participating in a RCT – which may limit the

338

generalizability of our findings. The assessment of PUFA exposure at only one-time point cannot

339

capture PUFA status changes throughout follow-up, and the aforementioned need to remediate

340

the RBC PUFA data22 added variability to the exposure marker, diminishing our ability to

341

separate signal from noise. The two major factors that determine n-3 and n-6 PUFA levels –

AC C

333

12

ACCEPTED MANUSCRIPT

dietary PUFA intake and variants in FA desaturase and elongase genes - were not explored here,

343

in part because they were beyond the scope of this study, but also because of their own inherent

344

limitations. Estimated PUFA intakes derived from food records are inferior to circulating PUFA

345

levels as indicators of PUFA exposure64, 65, and the effects of genetic variants on PUFA levels,

346

while substantial for ARA66, are minimal to non-existent for EPA and DHA67. Finally, the

347

inability to rule out the possibility of residual or unmeasured confounding also precludes

348

inferences about causality68.

RI PT

342

349

In conclusion, we found that higher RBC levels of n-3 PUFAs were associated with greater

351

longevity in post-menopausal women. These findings, in the context of the totality of available

352

evidence on this subject, provide support for a role of n-3 PUFAs in reducing the risk for death

353

across a spectrum of causes.

AC C

EP

TE D

354 355

M AN U

SC

350

13

ACCEPTED MANUSCRIPT

ACKNOWLEDGEMENTS

357

The authors wish to thank Jason Polreis (OmegaQuant Analytics, LLC) for performing the RBC

358

fatty acid analyses. Author Contributions: WSH wrote the original grant that supported the

359

laboratory analysis of the blood samples and wrote the first draft of the manuscript. JL and JPV

360

analyzed the data and assisted with manuscript preparation; MAE and JGR collaborated on the

361

original grant submission and reviewed/edited the manuscript. KLM, JEM, LW and TMB

362

provided substantive input for manuscript revisions. All authors have read and approved the final

363

manuscript.

RI PT

356

SC

364

AC C

EP

TE D

M AN U

365

14

ACCEPTED MANUSCRIPT

References

6.

7. 8.

9.

10.

11. 12.

13.

14.

15. 16.

17.

18.

RI PT

SC

5.

M AN U

4.

TE D

3.

EP

2.

Itomura M, Fujioka S, Hamazaki K, et al. Factors influencing EPA+DHA levels in red blood cells in Japan. In Vivo. 2008;22:131-135. Ogura T, Takada H, Okuno M, et al. Fatty acid composition of plasma, erythrocytes and adipose: their correlations and effects of age and sex. Lipids. 2010;45:137-144. CIA. Country Comparison of Life Expectancy at Birth. 2011. https://www.cia.gov/library/publications/the-world-factbook/rankorder/2102rank.html. Pottala JV, Garg S, Cohen BE, Whooley MA, Harris WS. Blood Eicosapentaenoic and Docosahexaenoic Acids Predict All-Cause Mortality in Patients With Stable Coronary Heart Disease: The Heart and Soul Study. Circ.Cardiovasc.Qual.Outcomes. 2010;3:406-412. Harris WS, von Schacky C. The Omega-3 Index: a new risk factor for death from coronary heart disease? Prev Med. 2004;39:212-220. Farzaneh-Far R, Lin J, Epel ES, Harris WS, Blackburn EH, Whooley MA. Association of marine omega-3 fatty acid levels with telomeric aging in patients with coronary heart disease. JAMA. 2010;303:250-257. Wu JH, Lemaitre RN, King IB, et al. Circulating omega-6 polyunsaturated fatty acids and total and cause-specific mortality: the Cardiovascular Health Study. Circulation. 2014;130:1245-1253. Warensjo E, Sundstrom J, Vessby B, Cederholm T, Riserus U. Markers of dietary fat quality and fatty acid desaturation as predictors of total and cardiovascular mortality: a population-based prospective study. Am J Clin Nutr.2008.Jul.;88.(1):203.-9. 2008;88:203-209. Laaksonen DE, Nyyssonen K, Niskanen L, Rissanen TH, Salonen JT. Prediction of cardiovascular mortality in middle-aged men by dietary and serum linoleic and polyunsaturated fatty acids. Arch.Intern.Med. 2005;165:193-199. Mozaffarian D, Lemaitre RN, King IB, et al. Plasma phospholipid long-chain omega-3 fatty acids and total and cause-specific mortality in older adults: a cohort study. Annals of internal medicine. 2013;158:515-525. Lindberg M, Saltvedt I, Sletvold O, Bjerve KS. Long-chain n-3 fatty acids and mortality in elderly patients. Am J Clin Nutr. 2008;88:722-729. Harris WS, Kennedy KF, O'Keefe JH, Jr., Spertus JA. Red blood cell fatty acid levels improve GRACE score prediction of 2-yr mortality in patients with myocardial infarction. International journal of cardiology. 2013;168:53-59. Chien KL, Lin HJ, Hsu HC, et al. Comparison of predictive performance of various fatty acids for the risk of cardiovascular disease events and all-cause deaths in a community-based cohort. Atherosclerosis. 2013;230:140-147. Marklund M, Leander K, Vikstrom M, et al. Polyunsaturated Fat Intake Estimated by Circulating Biomarkers and Risk of Cardiovascular Disease and All-Cause Mortality in a Population-Based Cohort of 60-Year-Old Men and Women. Circulation. 2015;132:586-594. Eide IA, Jenssen T, Hartmann A, et al. The association between marine n-3 polyunsaturated fatty acid levels and survival after renal transplantation. Clin J Am Soc Nephrol. 2015;10:1246-1256. Kleber ME, Delgado GE, Lorkowski S, Marz W, von Schacky C. Omega-3 fatty acids and mortality in patients referred for coronary angiography. The Ludwigshafen Risk and Cardiovascular Health Study. Atherosclerosis. 2016;252:175-181. Pottala JV, Yaffe K, Robinson JG, Espeland MA, Wallace R, Harris WS. Higher RBC EPA + DHA corresponds with larger total brain and hippocampal volumes: WHIMS-MRI Study. Neurology. 2014;82:435-442. Ammann EM, Pottala JV, Harris WS, et al. Omega-3 fatty acids and domain-specific cognitive aging: Secondary analyses of data from WHISCA. Neurology. 2013;81:1484-1491.

AC C

1.

15

ACCEPTED MANUSCRIPT

24.

25.

26. 27.

28.

29. 30. 31.

32. 33.

34.

35.

36.

RI PT

SC

23.

M AN U

22.

TE D

21.

EP

20.

Persons JE, Robinson JG, Ammann EM, et al. Omega-3 fatty acid biomarkers and subsequent depressive symptoms. Int J Geriatr Psychiatry. 2013. Bazinet RP, Chu MW. Omega-6 polyunsaturated fatty acids: is a broad cholesterol-lowering health claim appropriate? CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2014;186:434-439. Shumaker SA, Reboussin BA, Espeland MA, et al. The Women's Health Initiative Memory Study (WHIMS): a trial of the effect of estrogen therapy in preventing and slowing the progression of dementia. Controlled clinical trials. 1998;19:604-621. Pottala JV, Espeland MA, Polreis J, Robinson J, Harris WS. Correcting the Effects of -20 degrees C Storage and Aliquot Size on Erythrocyte Fatty Acid Content in the Women's Health Initiative. Lipids. 2012;47:835-846. Curb JD, McTiernan A, Heckbert SR, et al. Outcomes ascertainment and adjudication methods in the Women's Health Initiative. Annals of epidemiology. 2003;13:S122-128. Pottala JV, Djira GD, Espeland MA, Ye J, Larson MG, Harris WS. Structural equation modeling for analyzing erythrocyte fatty acids in Framingham. Computational and mathematical methods in medicine. 2014:ID:160520. Browning LM, Walker CG, Mander AP, et al. Incorporation of eicosapentaenoic and docosahexaenoic acids into lipid pools when given as supplements providing doses equivalent to typical intakes of oily fish. Am J Clin.Nutr. 2012;96:748-758. Del Gobbo LC, Imamura F, Aslibekyan S, et al. omega-3 Polyunsaturated Fatty Acid Biomarkers and Coronary Heart Disease: Pooling Project of 19 Cohort Studies. JAMA internal medicine. 2016. Mozaffarian D, Lemaitre RN, King IB, et al. Circulating long-chain omega-3 fatty acids and incidence of congestive heart failure in older adults: the cardiovascular health study: a cohort study. Annals of internal medicine. 2011;155:160-170. Sala-Vila A, Guasch-Ferre M, Hu FB, et al. Dietary alpha-Linolenic Acid, Marine omega-3 Fatty Acids, and Mortality in a Population With High Fish Consumption: Findings From the PREvencion con DIeta MEDiterranea (PREDIMED) Study. Journal of the American Heart Association. 2016;5. Harris WS, Poston WC, Haddock CK. Tissue n-3 and n-6 fatty acids and risk for coronary heart disease events. Atherosclerosis. 2007;193:1-10. Harris WS. The omega-6/omega-3 ratio and cardiovascular disease risk: uses and abuses. Curr Atheroscler Rep. 2006;8:453-459. Chowdhury R, Warnakula S, Kunutsor S, et al. Association of dietary, circulating, and supplement fatty acids with coronary risk: a systematic review and meta-analysis. Annals of internal medicine. 2014;160:398-406. Folsom AR, Demissie Z. Fish intake, marine omega-3 fatty acids, and mortality in a cohort of postmenopausal women. American journal of epidemiology. 2004;160:1005-1010. Engeset D, Braaten T, Teucher B, et al. Fish consumption and mortality in the European Prospective Investigation into Cancer and Nutrition cohort. European journal of epidemiology. 2015;30:57-70. Bell GA, Kantor ED, Lampe JW, Kristal AR, Heckbert SR, White E. Intake of long-chain omega-3 fatty acids from diet and supplements in relation to mortality. American journal of epidemiology. 2014;179:710-720. Balk EM, Lichtenstein AH, Chung M, Kupelnick B, Chew P, Lau J. Effects of omega-3 fatty acids on serum markers of cardiovascular disease risk: a systematic review. Atherosclerosis. 2006;189:1930. Geleijnse JM, Giltay EJ, Grobbee DE, Donders AR, Kok FJ. Blood pressure response to fish oil supplementation: metaregression analysis of randomized trials. J.Hypertens. 2002;20:1493-1499.

AC C

19.

16

ACCEPTED MANUSCRIPT

42.

43. 44. 45.

46.

47.

48.

49. 50.

51. 52. 53.

RI PT

SC

41.

M AN U

40.

TE D

39.

EP

38.

Violi F, Pignatelli P, Basili S. Nutrition, supplements, and vitamins in platelet function and bleeding. Circulation. 2010;121:1033-1044. Mozaffarian D, Geelen A, Brouwer IA, Geleijnse JM, Zock PL, Katan MB. Effect of fish oil on heart rate in humans: a meta-analysis of randomized controlled trials. Circulation. 2005;112:19451952. Rauch B, Senges J. The effects of supplementation with omega-3 polyunsaturated Fatty acids on cardiac rhythm: anti-arrhythmic, pro-arrhythmic, both or neither? It depends... Frontiers in physiology. 2012;3:1-9. Fontes JD, Rahman F, Lacey S, et al. Red blood cell fatty acids and biomarkers of inflammation: A cross-sectional study in a community-based cohort. Atherosclerosis. 2015;240:431-436. Li K, Huang T, Zheng J, Wu K, Li D. Effect of marine-derived n-3 polyunsaturated fatty acids on Creactive protein, interleukin 6 and tumor necrosis factor alpha: a meta-analysis. PloS one. 2014;9:e88103. Cawood AL, Ding R, Napper FL, et al. Eicosapentaenoic acid (EPA) from highly concentrated n-3 fatty acid ethyl esters is incorporated into advanced atherosclerotic plaques and higher plaque EPA is associated with decreased plaque inflammation and increased stability. Atherosclerosis. 2010;212:252-259. Thies F, Garry JM, Yaqoob P, et al. Association of n-3 polyunsaturated fatty acids with stability of atherosclerotic plaques: a randomised controlled trial. Lancet. 2003;361:477-485. Pase MP, Grima NA, Sarris J. Do long-chain n-3 fatty acids reduce arterial stiffness? A metaanalysis of randomised controlled trials. The British journal of nutrition. 2011;106:974-980. Flock MR, Skulas-Ray AC, Harris WS, Etherton TD, Fleming JA, Kris-Etherton PM. Determinants of Erythrocyte Omega-3 Fatty Acid Content in Response to Fish Oil Supplementation: A DoseResponse Randomized Controlled Trial. Journal of the American Heart Association. 2013;2:e000513. Iso H, Kobayashi M, Ishihara J, et al. Intake of fish and n3 fatty acids and risk of coronary heart disease among Japanese: the Japan Public Health Center-Based (JPHC) Study Cohort I. Circulation. 2006;113:195-202. Investigators G-H. Effect of n-3 polyunsaturated fatty acids in patients with chronic heart failure (the GISSI-HF trial): a randomised, double-blind, placebo-controlled trial. Lancet. 2008;372:12231230. Marchioli R, Barzi F, Bomba E, et al. Early protection against sudden death by n-3 polyunsaturated fatty acids after myocardial infarction: time-course analysis of the results of the Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico (GISSI)-Prevenzione. Circulation. 2002;105:1897-1903. Burr ML, Fehily AM, Gilbert JF, et al. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: diet and reinfarction trial (DART). Lancet. 1989;2:757-761. Rauch B, Schiele R, Schneider S, et al. OMEGA, a randomized, placebo-controlled trial to test the effect of highly purified omega-3 fatty acids on top of modern guideline-adjusted therapy after myocardial infarction. Circulation. 2010;122:2152-2159. Bosch J, Gerstein HC, Dagenais GR, et al. n-3 fatty acids and cardiovascular outcomes in patients with dysglycemia. The New England journal of medicine. 2012;367:309-318. Roncaglioni MC, Tombesi M, Avanzini F, et al. n-3 fatty acids in patients with multiple cardiovascular risk factors. The New England journal of medicine. 2013;368:1800-1808. Harris WS, Sands SA, Windsor SL, et al. Omega-3 Fatty Acids in Cardiac Biopsies from Heart Transplant Patients: Correlation with Erythrocytes and Response to Supplementation. Circulation. 2004;110:1645-1649.

AC C

37.

17

ACCEPTED MANUSCRIPT

59. 60. 61.

62. 63. 64.

65.

66.

67.

68.

RI PT

SC

58.

M AN U

57.

TE D

56.

EP

55.

Gurzell EA, Wiesinger JA, Morkam C, Hemmrich S, Harris WS, Fenton JI. Is the omega-3 index a valid marker of intestinal membrane phospholipid EPA+DHA content? Prostaglandins, leukotrienes, and essential fatty acids. 2014. Metcalf RG, Cleland LG, Gibson RA, et al. Relation between blood and atrial fatty acids in patients undergoing cardiac bypass surgery. Am J Clin Nutr. 2010;91:528-534. Rauch B, Schiele R, Schneider S, et al. Highly purified omega-3 fatty acids for secondary prevention of sudden cardiac death after myocardial infarction-aims and methods of the OMEGA-study. Cardiovasc.Drugs Ther. 2010;20:365-375. Galan P, Kesse-Guyot E, Czernichow S, Briancon S, Blacher J, Hercberg S. Effects of B vitamins and omega 3 fatty acids on cardiovascular diseases: a randomised placebo controlled trial. BMJ. 2010;341:c6273. Eussen SR, Geleijnse JM, Giltay EJ, Rompelberg CJ, Klungel OH, Kromhout D. Effects of n-3 fatty acids on major cardiovascular events in statin users and non-users with a history of myocardial infarction. Eur Heart J. 2012;33:1582-1588. Harris WS. Are n-3 fatty acids still cardioprotective? Current opinion in clinical nutrition and metabolic care. 2013;16:141-149. Deckelbaum RJ, Calder PC. Different outcomes for omega-3 heart trials: why? Current opinion in clinical nutrition and metabolic care. 2012;15:97-98. Manson JE, Bassuk SS, Lee IM, et al. The VITamin D and OmegA-3 TriaL (VITAL): rationale and design of a large randomized controlled trial of vitamin D and marine omega-3 fatty acid supplements for the primary prevention of cancer and cardiovascular disease. Contemp Clin Trials. 2012;33:159-171. Bowen KJ, Harris WS, Kris-Etherton PM. Omega-3 Fatty Acids and Cardiovascular Disease: Are There Benefits? Current treatment options in cardiovascular medicine. 2016;18:69. Harris WS, Thomas RM. Biological variability of blood omega-3 biomarkers. Clinical biochemistry. 2010;43:338-340. Virtanen JK, Mursu J, Voutilainen S, Uusitupa M, Tuomainen TP. Serum omega-3 polyunsaturated Fatty acids and risk of incident type 2 diabetes in men: the kuopio ischemic heart disease risk factor study. Diabetes Care. 2014;37:189-196. de Oliveira Otto MC, Wu JH, Baylin A, et al. Circulating and Dietary Omega-3 and Omega-6 Polyunsaturated Fatty Acids and Incidence of CVD in the Multi-Ethnic Study of Atherosclerosis. Journal of the American Heart Association. 2013;2:e000506. Zhang JY, Kothapalli KS, Brenna JT. Desaturase and elongase-limiting endogenous long-chain polyunsaturated fatty acid biosynthesis. Current opinion in clinical nutrition and metabolic care. 2016;19:103-110. Tintle NL, Pottala JV, Lacey S, et al. A genome-wide association study of saturated, mono- and polyunsaturated red blood cell fatty acids in the Framingham Heart Offspring Study. Prostaglandins, leukotrienes, and essential fatty acids. 2015;94:65-72. Harris WS, Kennedy KF, Maddox TM, Kutty S, Spertus JA. Multiple differences between patients who initiate fish oil supplementation post-myocardial infarction and those who do not: the TRIUMPH Study. Nutr Res. 2016;36:65-71.

AC C

54.

18

ACCEPTED MANUSCRIPT

Table 1. Baseline characteristics by quartile of the Omega-3 Index.

305 (19.3) 351 (22.2) 478 (30.3) 445 (28.1)

345 (21.6) 320 (20.0) 470 (29.4) 462 (28.9)

347 (20.5) 315 (18.6) 523 (30.9) 507 (30.09)

279 (17.1) 289 (17.7) 496 (30.4) 569 (34.8)

2 (0.1) 42 (2.7) 55 (3.5) 1457 (92.3) 23 (1.5)

14 (0.9) 80 (5.0) 41 (2.6) 1443 (90.4) 19 (1.2)

18 (1.1) 132 (7.8) 21 (1.2) 1490 (88.1) 31 (1.8)

83 (5.1) 180 (11.0) 23 (1.4) 1307 (80.0) 40 (2.5)

561 (35.5) 658 (41.7) 221 (14.0) 139 (8.8)

530 (33.2) 672 (42.1) 267 (16.7) 128 (8.0)

441 (26.1) 718 (42.4) 304 (18.0) 229 (13.5)

389 (23.8) 580 (35.5) 348 (21.3) 316 (19.4)

432 (27.1) 558 (34.9) 380 (23.8) 163 (10.2) 64 (4.0) 89.7 (13.6) 10.6 (12.4) 167 (10.5)

462 (27.3) 675 (39.9) 354 (20.9) 146 (8.6) 55 (3.3) 88.6 (13.0) 11.5 (13.0) 218 (12.9)

581 (35.6) 606 (37.1) 320 (19.6) 92 (5.6) 34 (2.1) 86.1 (12.4) 14.1 (15.2) 248 (15.2)

<0.0001 <0.0001

440 (27.9) 563 (35.7) 356 (22.6) 150 (9.5) 70 (4.4) 89.4 (13.8) 9.3 (12.3) 153 (9.7)

SC

<0.0001

888 (56.2) 216 (13.7) 121 (7.7) 92 (5.8) 56 (3.6) 206 (13.1) 1015 (64.3) 1105 (67.0)

890 (55.7) 236 (14.8) 112 (7.0) 111 (7.0) 58 (3.6) 190 (11.9) 1051 (65.8) 1136 (71.1)

913 (54.0) 285 (16.8) 138 (8.2) 119 (7.0) 82 (4.9) 155 (9.2) 1094 (64.7) 1171 (69.2)

822 (50.3) 309 (18.9) 168 (10.3) 117 (7.2) 65 (4.0) 152 (9.3) 1013 (62.0) 1112 (68.1)

304 (19.3) 351 (22.2) 548 (34.7) 376 (23.8) 394 (25.0) 236 (15.0) 776 (49.2) 95 (6.0) 261 (16.5) 63 (4.0)

360 (22.5) 364 (22.8) 445 (27.9) 428 (26.8) 430 (26.9) 272 (17.0) 829 (51.9) 121 (7.6) 276 (17.3) 63 (3.9)

481 (28.4) 356 (21.0) 377 (22.3) 478 (28.3) 460 (27.2) 288 (17.0) 796 (47.0) 109 (6.4) 304 (18.0) 60 (3.6)

654 (40.1) 279 (17.1) 204 (12.5) 496 (30.4) 437 (26.8) 361 (22.1) 812 (49.7) 95 (5.8) 282 (17.3) 48 (2.9)

EP

AC C

1

P value1

RI PT

4th 1633 7.11% (6.55, 8.05) 70.4 (3.8)

TE D

Age at baseline [years]: mean (SD) HT randomization arms: n (%) Estrogen (E) - intervention E - control E+Progesterone (P) intervention E+P control Race: n (%) Asian or Pacific Islander Black or African American Hispanic or Latino Non-Hispanic White Other Highest Education: n (%) High School or Less Some College/Tech School Bachelor’s Degree Master’s Degree or Higher Body Mass Index [kg/m2]: n (%) < 25 [25 – 30) [30 – 35) [35 – 40) ≥ 40 Waist circumference [cm]: mean (SD) Physical activity [METs/wk]: mean (SD) Alcohol >7 drinks/wk: n (%) Smoking pack-years: n (%) Never smoking <10 [10 – 20) [20 - 30) [30 – 40) ≥ 40 Family history of cancer: n (%) Family history of CVD: n (%) Region of USA Northeast South Midwest West Aspirin use: n (%) Taking pills for cholesterol: n (%) Hypertension: n (%) Diabetes: n (%) Cardiovascular disease2: n (%) History of Cancer: n (%)

Omega-3 index (quartile) 2nd 3rd 1597 1692 4.59% 5.53% (4.39, 4.81) (5.27, 5.80) 70.0 (3.8) 70.3 (3.8)

M AN U

N Omega-3 Index (median, IQR)

1st 1579 3.59% (3.14, 3.88) 69.8 (3.9)

<0.0001

<0.0001

<0.0001 <0.0001 <0.0001 <0.0001

0.2 0.3 <0.0001

0.5 <0.0001 0.048 0.2 0.8 0.3

A Chi-squared test was used to evaluate differences for categorical variables, and a t-test or one-way ANOVA was

used for continuous variables.

2

Yes to the question, “Has a doctor ever told you that you had heart problems,

problems with your blood circulation, or blood clots?”

19

ACCEPTED MANUSCRIPT

Table 2. RBC fatty acids (percent of total) by mortality status at last contact (medians and interquartile ranges). Alive (n=4650)

Dead (n=1851) 25th

75th

P-value

6.10

4.92

4.13

5.89

0.0003

0.50

0.83

0.61

0.48

0.78

<0.0001

4.40

3.68

5.31

4.31

3.61

5.14

0.0016

Linoleic

11.89

10.87

12.98

11.84

10.74

12.91

0.15

PUFA Factor Score*

-6.75

-6.98

-6.49

-6.81

-7.03

-6.56

<0.0001

Alpha-linolenic

0.16

0.13

0.20

0.16

0.13

0.19

0.08

Docosapentaenoic n-3 (DPAn-3)

2.55

2.31

2.79

2.49

2.26

2.74

<0.0001

Arachidonic (ARA)

17.06

15.94

18.08

17.03

15.93

18.12

0.98

ARA/EPA ratio

27.04

20.34

34.05

28.55

21.82

35.52

<0.0001

N-6/N-3 ratio

6.88

5.89

7.85

7.03

6.05

7.97

<0.0001

Omega-3 index (EPA+DHA)

5.04

4.25

Eicosapentaenoic (EPA)

0.64

Docosahexaenoic (DHA)

75

M AN U

25

RI PT

Median

Median

th

SC

th

*PUFA Factor score is a continuous metric derived as follows: The latter included a weighted

combination of six RBC PUFAs: 0.37*lnALA + 0.84*lnEPA + 0.80*DHA - 0.667*ARA 0.814*adrenic - 0.788*DPAn-624. Higher levels reflect more n-3 PUFAs and/or less n-6 PUFAs. Wilcoxon Rank Sums test was used to test for differences between groups. Critical levels for the primary

AC C

EP

TE D

PUFA metrics, p≤0.01. Significant p-values are shown in bold.

20

ACCEPTED MANUSCRIPT

Table 3. Multivariable-adjusted hazard ratios (99% CI) per 1-SD increase in red blood cell fatty acids for total mortality over 14.9 (median) years follow up from the fully adjusted modela. Omega-3 index (EPA+DHA) Eicosapentaenoic (EPA) Docosahexaenoic (DHA) Linoleic Acid PUFA Factor Scorec Alpha-linolenic Docosapentaenoic n-3 Arachidonic (ARA) ARA/EPA ratio N-6/N-3 ratio

0.92 (0.85, 0.98) 0.89 (0.82, 0.96) 0.93 (0.87, 1.00) 0.99 (0.93, 1.06) 0.77 (0.66, 0.91) 0.98 (0.91, 1.05) 0.93 (0.86, 1.00) 1.00 (0.92, 1.08) 1.12 (1.03, 1.21) 1.10 (1.02, 1.19)

SC

0.92 (0.85, 0.98) 0.88 (0.82, 0.95) 0.93 (0.87, 1.00) 0.99 (0.92, 1.06) 0.77 (0.65, 0.91) 0.98 (0.91, 1.05) 0.93 (0.86, 1.00) 1.00 (0.92, 1.08) 1.12 (1.03, 1.22) 1.10 (1.02, 1.19)

Model 3

M AN U

a

0.90 (0.84, 0.96) 0.87 (0.81, 0.93) 0.91 (0.85, 0.98) 0.97 (0.91, 1.04) 0.74 (0.63, 0.86) 0.97 (0.9, 1.04) 0.91 (0.84, 0.97) 1.00 (0.91, 1.09) 1.14 (1.06, 1.23) 1.13 (1.04, 1.22)

Model 2

P-valueb 0.0015 <.0001 0.0099 0.8286 <.0001 0.3689 0.014 0.8956 0.0004 0.0021

RI PT

Model 1

Model 1 adjusted for age, race, HT assignment. Model 2 added BMI, highest education, smoking pack-

year, physical activity, weekly alcohol intake, waist circumference, region, family history of cancer, family history of CVD, and aspirin use. Model 3 (fully adjusted) added high cholesterol requiring pills (ever), and a history of hypertension, diabetes, cardiovascular disease and/or cancer. b For model 3; critical values for the Omega-3 Index, EPA and DHA were set a priori at <0.02; for all other metrics in

AC C

EP

TE D

the Table, <0.01. P-values in bold are significant. c See Table 2.

21

ACCEPTED MANUSCRIPT

Figure Legends

Figure 1. Cumulative Nelson-Aalen unadjusted hazard estimates of total mortality by quartiles of baseline omega-3 index (left) and PUFA Factor score (right). Log Rank p-value <0.001 for

RI PT

both. The numbers of participants at risk at each biennial examination are shown.

Figure 2. Fully Adjusted HRs for Total Mortality by Quartile of Selected RBC PUFA. Error bars

AC C

EP

TE D

M AN U

SC

are 95% CIs, and p-values for trends are given above the columns.

22

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Highlights Relations between RBC fatty acids and total mortality were studied in 6501 women. Multi-variable adjusted odds ratios per 1-SD increase in FA levels were calculated.

RI PT

There was an 8% reduction in risk associated with higher EPA+DHA levels.

AC C

EP

TE D

M AN U

SC

RBC omega-3 FA content is a significant predictor of all-cause mortality.

ACCEPTED MANUSCRIPT

Supplemental Table 1. Selected studies linking total mortality with n-3 PUFAs (from dietary records, supplementation use, randomized trials or circulating biomarkers). Patient Type

Years Followup

Results

Dietary Data Bell(1)

70,495

5

Yamagishi(2)

57,972

Vitamins and Lifestyle Study, USA Healthy Japanese

12.7

Yuan(3) Folsom(4)

18,224 41,836

Healthy Chinese men Iowa Women’s Study

12 14

Engeset(5)

480,535

11-18

Sala-Vila(6)

7202

European Prospective Investigation into Cancer and Nutrition (EPIC) PREDIMED Cohort

Reduced total (and cancer) mortality associated with higher LC n-3PUFA intake Trend (p=0.1) for lower mortality with higher LC n-PUFA intakes (1.0 to 2.3 g/d) Fish intake inversely associated with mortality No associations with mortality across n-3 FA intake of 20 to 470 mg/d No associations between fish intake and total mortality

EP

TE D

M AN U

Fish Oil Use – Non-randomized trials Poole(7) 12,178 Immediately Post-MI Prescribed FOS or not Patel(8) 3827 Hyperlipidemic patients prescribed statins with or without ezetimibe Macchia(9) 11,532 Post MI patients taking statins with or without fish oil Fish Oil Use – Randomized trials Rizos(10) 63,279 Meta-analysis of 17 trials with fish oils Casula(11) 15,384 Meta-analysis of 11 secondary prevention trials giving at least 1 g n-3 FA for at least 1 year Biomarker Wu(12) 2792 Generally healthy elderly Plasma (mean age 74 at baseline) phospholipids (PL) Warensjo(13) 279 cases Population-based cohort Plasma 279 controls cholesteryl esters (CE) Laaksonen(14) 1551 Population-based cohort of Plasma men Mozaffarian(15) 2692 Generally healthy elderly Plasma PL (mean age 74 at baseline) Lindberg(16) 254 Elderly admitted for acute Plasma PL illness Harris(17) 1114 Immediately Post-MI Red Blood Cells (RBC) Pottala(18) 956 History of CVD Whole Blood Chien(19) 1833 Healthy Taiwanese Plasma Marklund(20) 4232 Generally healthy cohort Plasma CE (mean age 60)

Patients meeting intake recommendation for ALA had reduced total mortality, with greatest reduction in those meeting recommendations for ALA and EPA+DHA

SC

5.9

RI PT

N

AC C

1 2

1

2.6

3

Prescription of FOS associated with reduced total mortality Only predictors of reduced mortality were higher HDL-C and reported use of fish oil supplements

4

Significant reduction in mortality in those on combination therapy

2.65 (mean) 2.0 (mean)

No reduction in total mortality with fish oil use [HR = 0.96 (95% CI, 0.91-1.02)] No reduction in total mortality with fish oil use [HR = 0.89 (95% CI, 0.78-1.02)] but significant reductions in cardiac and sudden death and myocardial infarction

11.5

Higher plasma PL LA associated with lower risk for total (and CVD) mortality. Higher levels of total n-6 and n-3 associated with the lowest total mortality risk Only PUFA associated with total mortality was LA (inversely)

12.5

15 11.5 3 2

5.6 9.6 14.9

Serum LA associated with lower total mortality; data for EPA/DHA not reported Higher plasma PL LC n-3 PUFA associated with lower total and CVD mortality Higher plasma PL EPA (not DHA) associated with reduced mortality (Q1 v Q234) Higher EPA associated with lower mortality

Reduced mortality associated with higher blood EPA+DHA Plasma EPA inversely associated with total mortality EPA and DHA (and LA) inversely associated with total mortality

ACCEPTED MANUSCRIPT

Eide(21) Plasma PL Kleber(22) RBC Current Study RBC

1990 3259 6501

Renal transplant recipients in Norway Suspected CAD

6.8

Generally healthy postmenopausal women (mean age 70)

14.9

10

Higher marine n-3 PUFA associated with lower total and CVD mortality Higher marine n-3 PUFA associated with lower total mortality Higher Omega-3 Index, EPA and DHA associated with reduced total mortality; EPA also associated with reduced CVD mortality

RI PT

3

AC C

EP

TE D

M AN U

SC

4

2

ACCEPTED MANUSCRIPT

Supplementaal Table 2. Participant’s baseline characteristics by survival status at last contact and causes of death during 14.9-years median follow up (IQR (12.9-15.9) Alive

Dead P value1 1851 71.5 (4.0)

909 (19.6) 892 (19.2) 1408 (30.3) 1441 (31.0)

367 (19.8) 383 (20.7) 559 (30.2) 542 (29.3)

90 (1.9) 328 (7.1) 120 (2.6) 4025 (87) 87 (1.9)

27 (2.0) 106 (5.7) 20 (1.1) 1672 (90.3) 26 (1.4)

1373 (29.5) 1848 (39.7) 832 (17.9) 597 (12.8)

548 (29.6) 780 (42.1) 308 (16.6) 215 (11.6)

Cause of death CVD Others 617 772 72.0 (4.1) 71.7 (3.9) 133 (21.6) 147 (23.8) 177 (28.7) 160 (25.9)

154 (20.0) 143 (18.5) 232 (30.1) 243 (31.5)

8 (1.7) 32 (6.9) 4 (0.9) 411 (89.0) 7 (1.5)

11 (1.8) 43 (7.0) 4 (0.7) 552 (89.5) 7 (1.1)

8 (1.0) 31 (1.6) 12 (1.6) 709 (91.8) 12 (1.6)

137 (30.0) 204 (44.2) 73 (15.8) 48 (10.4)

176 (28.5) 277 (44.9) 89 (14.4) 75 (12.2)

235 (30.4) 299 (38.7) 146 (18.9) 92 (12.5)

140 (30.3) 154 (33.3) 112 (24.2) 47 (10.2) 9 (2.0) 89.8 (13.5) 10.2 (12.0) 74 (16.0)

171 (27.7) 231 (37.4) 139 (22.5) 53 (8.6) 23 (3.7) 90.0 (13.4) 9.5 (12.0) 69 (11.2)

248 (32.1) 287 (37.2) 150 (19.4) 59 (7.6) 28 (3.6) 88.7 (13.6) 10.2 (14.0) 101 (13.1)

180 (39.0) 66 (14.3) 35 (7.6) 49 (10.6) 24 (5.2) 108 (23.4) 313 (67.8) 315 (68.2)

306 (49.6) 102 (16.5) 44 (7.1) 41 (6.7) 25 (4.1) 99(16.1) 231 (62.6) 452 (73.3)

382 (49.5) 105 (13.6) 54 (7.0) 61 (7.9) 33 (4.3) 137 (17.8) 302 (60.9) 535 (69.3)

140 (30.3) 75 (16.2) 107 (23.2) 140 (30.3) 133 (28.8) 76 (16.5) 226 (48.9) 30 (6.5) 90 (19.5) 25 (5.4)

172 (27.9) 135 (21.9) 131 (21.2) 179 (29.0) 183 (29.7) 142 (23.0) 410 (66.5) 89 (14.4) 194 (31.4) 29 (4.7)

22 (28.8) 169 (21.9) 182 (23.6) 199 (25.8) 192 (24.9) 128 (16.6) 435 (56.4) 73 (9.5) 162 (21.0) 33 (4.3)

SC

80 (17.3) 93 (20.1) 150 (32.5) 139 (30.1)

M AN U

0.2

559 (30.2) 672 (36.3) 401 (21.7) 159 (8.6) 60 (3.2) 89.4 (13.5) 10.0 (12.9) 244 (13.2)

2645 (56.9) 773 (16.6) 406 (8.7) 288 (6.2) 179 (3.9) 359 (7.7) 3004 (64.6) 3222 (69.3)

868 (46.9) 273 (14.8) 133 (7.2) 151 (8.2) 82 (4.4) 344 (18.6) 1169 (63.2) 1302 (70.3)

1265 (27.2) 971 (20.9) 1154 (24.8) 1260 (27.1) 1213 (26.1) 811 (17.4) 2142 (46.1) 228 (4.9) 677 (14.6) 147 (3.2)

534 (28.9) 379 (20.5) 420 (22.7) 518 (28.0) 508 (27.4) 346 (18.7) 1071 (57.9) 192 (10.4) 446 (24.1) 87 (4.7)

8

1

9

ANOVA was used for continuous variables.

0.0002 <0.0001 0.09 <0.0001

0.3 0.4 0.2

0.3 0.2 <0.0001 <0.0001 <0.0001 0.0034

3

<0.0001 0.06

0.1

0.1

A Chi-squared test was used to evaluate differences for categorical variables, and a t-test or one-way

10

P value1

0.2

0.9

1356 (29.2) 1730 (37.2) 1009 (21.7) 392 (8.4) 163 (3.5) 88.0 (13.2) 11.9 (13.5) 542 (11.7)

EP

AC C

<0.0001 0.41

0.0001

TE D

Age at baseline [years]: mean (SD) HT randomization arms: n (%) E-alone intervention E-alone control E+P intervention E+P control Race: n (%) Asian or Pacific Islander Black or African American Hispanic or Latino Non-Hispanic White Other Highest Education: n (%) High School or Less Some College/Tech School Bachelor’s Degree Master’s Degree or Higher Body Mass Index [kg/m2]: n (%) < 25 [25 – 30) [30 – 35) [35 – 40) ≥ 40 Waist circumference [cm]: mean (SD) Physical activity [METs/wk]: mean (SD) Alcohol >7 drinks/wk: n (%) Smoking pack-years: n (%) Never smoking <10 [10 – 20) [20 - 30) [30 – 40) ≥ 40 Family history of cancer: n (%) Family history of CVD: n (%) Region of USA Northeast South Midwest West Aspirin use: n (%) Taking pills for cholesterol: n (%) Hypertension: n (%) Diabetes: n (%) Cardiovascular disease: n (%) History of Cancer: n (%)

4650 69.6 (3.6)

Cancer 462 70.7 (4.0)

RI PT

5 6 7

0.2 0.5 0.07 0.006

0.05 0.1 0.2

0.1 0.0034 <0.0001 <0.0001 <0.0001 0.7

ACCEPTED MANUSCRIPT

Supplemental Table 3. Sensitivity Analyses for Associations of RBC PUFA and Total Mortality: Multivariable-adjusted hazard ratios (99% CI) per 1-SD in the fully-adjusted modela Excluding death in first two years (n=6436)

Omega-3 index (EPA+DHA)

0.91 (0.85, 0.98)

Stratified by ASA use at baselineb

Excluding CVD/cancer at baseline (n=5144) 0.91 (0.84, 0.99)

Yes (n=1721)

0.9 (0.83, 0.99)

0.96 (0.83, 1.1)

0.88 (0.8, 0.95)

0.92 (0.79, 1.06)

0.92 (0.84, 1.01)

0.97 (0.84, 1.11)

1 (0.92, 1.08)

0.99 (0.87, 1.12)

0.75 (0.61, 0.91)

0.85 (0.64, 1.13)

0.97 (0.88, 1.07)

1 (0.86, 1.15)

0.92 (0.84, 1)

0.96 (0.84, 1.11)

0.99 (0.89, 1.1)

1.02 (0.85, 1.22)

1.13 (1.03, 1.24)

1.09 (0.93, 1.28)

1.12 (1.01, 1.24)

1.06 (0.92, 1.22)

13

M AN U

SC

Eicosapentaenoic (EPA) 0.89 (0.82, 0.96) 0.88 (0.8, 0.97) 0.93 (0.85, 1.01) Docosahexaenoic (DHA) 0.93 (0.86, 1) 0.99 (0.92, 1.06) 1.01 (0.93, 1.1) Linoleic Acid (LA) c PUFA Factor Score 0.77 (0.65, 0.91) 0.78 (0.63, 0.96) 0.97 (0.9, 1.05) 1 (0.91, 1.09) Alpha linolenic (ALA) 0.93 (0.86, 1.01) 0.93 (0.85, 1.02) Docosapentaenoic n-3 (DPAn-3) 1 (0.92, 1.09) 1.01 (0.91, 1.12) Arachidonic (ARA) ARA/EPA ratio 1.12 (1.03, 1.22) 1.13 (1.03, 1.25) N-6/N-3 ratio 1.1 (1.02, 1.2) 1.12 (1.02, 1.22) a Model components shown in Table 3. b Interactions all p>0.08. c See Table 2.

No (n=4780)

RI PT

11 12

AC C

EP

TE D

14

4

ACCEPTED MANUSCRIPT

Supplemental Table 4. Multivariable-adjusted hazard ratios (99% CI) per 1-SD increase in red blood cell polyunsaturated fatty acids for CVD, cancer and other mortality over 14.9 (median) years follow up from the fully adjusted modela.

0.97 (0.85, 1.12) 0.88 (0.77, 1.00) 1 (0.87, 1.14) 0.96 (0.86, 1.08) 0.76 (0.57, 1.01) 0.97 (0.84, 1.12) 0.92 (0.8, 1.05) 1 (0.85, 1.18) 1.13 (0.99, 1.29) 1.05 (0.9, 1.23)

Omega-3 index (EPA+DHA) Eicosapentaenoic (EPA) Docosahexaenoic (DHA) Linoleic Acid (LA) PUFA Factor Score Alpha linolenic (ALA) Docosapentaenoic n-3 (DPAn-3) Arachidonic (ARA) ARA/EPA ratio N-6/N-3 ratio a

0.92 (0.79, 1.07) 0.91 (0.78, 1.07) 0.93 (0.79, 1.09) 0.94 (0.83, 1.07) 0.81 (0.57, 1.16) 0.98 (0.85, 1.12) 0.93 (0.81, 1.07) 1.03 (0.89, 1.18) 1.1 (0.93, 1.3) 1.1 (0.93, 1.3)

M AN U

18 19

Cancer Death (n=462)

Model 3 described in legend for Table 2. Values in bold are significant (p<0.01).

AC C

EP

TE D

20

5

Other Death (n=772) 0.89 (0.78, 1.01) 0.87 (0.76, 1.01) 0.9 (0.79, 1.02) 1.05 (0.93, 1.17) 0.76 (0.55, 1.04) 0.98 (0.86, 1.13) 0.94 (0.82, 1.07) 0.98 (0.84, 1.14) 1.13 (0.97, 1.31) 1.13 (0.97, 1.31)

RI PT

CVD Death (n=617)

SC

15 16 17

ACCEPTED MANUSCRIPT

Supplemental Figure 1. Risk for death from any cause was significantly lower at an Omega-3

Index >8% vs <4% (HR, 0.69; 95% CI 0.52 to 0.93, p=0.017), and the trend across the 3 categories (<4%, 4-8% and >8%) was significant (p=0.019).

M AN U

SC

RI PT

21 22 23 24

25 26 27

3.

4. 5.

6.

TE D

2.

Bell GA, Kantor ED, Lampe JW, Kristal AR, Heckbert SR, White E. Intake of long-chain omega-3 fatty acids from diet and supplements in relation to mortality. American journal of epidemiology 2014;179(6):710-20. doi: 10.1093/aje/kwt326. Yamagishi K, Iso H, Date C, Fukui M, Wakai K, Kikuchi S, Inaba Y, Tanabe N, Tamakoshi A. Fish, omega-3 polyunsaturated fatty acids, and mortality from cardiovascular diseases in a nationwide community-based cohort of Japanese men and women the JACC (Japan Collaborative Cohort Study for Evaluation of Cancer Risk) Study. J Am CollCardiol2008Sep16;52(12):988-96 2008;52:988-96. Yuan JM, Ross RK, Gao YT, Yu MC. Fish and shellfish consumption in relation to death from myocardial infarction among men in Shanghai, China. American journal of epidemiology 2001;154:809-16. Folsom AR, Demissie Z. Fish intake, marine omega-3 fatty acids, and mortality in a cohort of postmenopausal women. American journal of epidemiology 2004;160:1005-10. Engeset D, Braaten T, Teucher B, Kuhn T, Bueno-de-Mesquita HB, Leenders M, Agudo A, Bergmann MM, Valanou E, Naska A, et al. Fish consumption and mortality in the European Prospective Investigation into Cancer and Nutrition cohort. European journal of epidemiology 2015;30(1):57-70. doi: 10.1007/s10654-014-9966-4. Sala-Vila A, Guasch-Ferre M, Hu FB, Sanchez-Tainta A, Bullo M, Serra-Mir M, Lopez-Sabater C, Sorli JV, Aros F, Fiol M, et al. Dietary alpha-Linolenic Acid, Marine omega-3 Fatty Acids, and Mortality in a Population With High Fish Consumption: Findings From the PREvencion con DIeta MEDiterranea (PREDIMED) Study. Journal of the American Heart Association 2016;5(1). doi: 10.1161/jaha.115.002543.

EP

1.

AC C

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

6

ACCEPTED MANUSCRIPT

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

RI PT

SC

10.

M AN U

9.

TE D

8.

Poole CD, Halcox JP, Jenkins-Jones S, Carr ES, Schifflers MG, Ray KK, Currie CJ. Omega-3 Fatty acids and mortality outcome in patients with and without type 2 diabetes after myocardial infarction: a retrospective, matched-cohort study. Clin Therap 2013;35(1):40-51. doi: 10.1016/j.clinthera.2012.11.008. Patel AY, Pillarisetti J, Marr J, Vacek JL. Ezetimibe in combination with a statin does not reduce all-cause mortality. Journal of clinical medicine research 2013;5(4):275-80. doi: 10.4021/jocmr1371w. Macchia A, Romero M, D'Ettorre A, Tognoni G, Mariani J. Exploratory analysis on the use of statins with or without n-3 PUFA and major events in patients discharged for acute myocardial infarction: an observational retrospective study. PloS one 2013;8(5):e62772. doi: 10.1371/journal.pone.0062772. Rizos EC, Ntzani EE, Bika E, Kostapanos MS, Elisaf MS. Association between omega-3 fatty acid supplementation and risk of major cardiovascular disease events: a systematic review and metaanalysis. JAMA 2012;308:1024-33. Casula M, Soranna D, Catapano AL, Corrao G. Long-term effect of high dose omega-3 fatty acid supplementation for secondary prevention of cardiovascular outcomes: A meta-analysis of randomized, double blind, placebo controlled trials. Atheroscler Suppl 2013;14(2):243-51. doi: 10.1016/s1567-5688(13)70005-9. Wu JH, Lemaitre RN, King IB, Song X, Psaty BM, Siscovick DS, Mozaffarian D. Circulating omega-6 polyunsaturated fatty acids and total and cause-specific mortality: the Cardiovascular Health Study. Circulation 2014;130(15):1245-53. doi: 10.1161/circulationaha.114.011590. Warensjo E, Sundstrom J, Vessby B, Cederholm T, Riserus U. Markers of dietary fat quality and fatty acid desaturation as predictors of total and cardiovascular mortality: a population-based prospective study. Am J Clin Nutr2008Jul;88(1):203-9 2008;88:203-9. Laaksonen DE, Nyyssonen K, Niskanen L, Rissanen TH, Salonen JT. Prediction of cardiovascular mortality in middle-aged men by dietary and serum linoleic and polyunsaturated fatty acids. ArchInternMed 2005;165:193-9. Mozaffarian D, Lemaitre RN, King IB, Song X, Huang H, Sacks FM, Rimm EB, Wang M, Siscovick DS. Plasma phospholipid long-chain omega-3 fatty acids and total and cause-specific mortality in older adults: a cohort study. Annals of internal medicine 2013;158(7):515-25. doi: 10.7326/0003-4819-158-7-201304020-00003. Lindberg M, Saltvedt I, Sletvold O, Bjerve KS. Long-chain n-3 fatty acids and mortality in elderly patients. Am J Clin Nutr 2008;88:722-9. Harris WS, Kennedy KF, O'Keefe JH, Jr., Spertus JA. Red blood cell fatty acid levels improve GRACE score prediction of 2-yr mortality in patients with myocardial infarction. International journal of cardiology 2013;168(1):53-9. doi: 10.1016/j.ijcard.2012.09.076. Pottala JV, Garg S, Cohen BE, Whooley MA, Harris WS. Blood Eicosapentaenoic and Docosahexaenoic Acids Predict All-Cause Mortality in Patients With Stable Coronary Heart Disease: The Heart and Soul Study. CircCardiovascQualOutcomes 2010;3:406-12. Chien KL, Lin HJ, Hsu HC, Chen PC, Su TC, Chen MF, Lee YT. Comparison of predictive performance of various fatty acids for the risk of cardiovascular disease events and all-cause deaths in a community-based cohort. Atherosclerosis 2013;230(1):140-7. doi: 10.1016/j.atherosclerosis.2013.06.015. Marklund M, Leander K, Vikstrom M, Laguzzi F, Gigante B, Sjogren P, Cederholm T, de Faire U, Hellenius ML, Riserus U. Polyunsaturated Fat Intake Estimated by Circulating Biomarkers and Risk of Cardiovascular Disease and All-Cause Mortality in a Population-Based Cohort of 60-YearOld Men and Women. Circulation 2015;132(7):586-94. doi: 10.1161/circulationaha.115.015607.

EP

7.

AC C

50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96

7

ACCEPTED MANUSCRIPT

21.

22.

Eide IA, Jenssen T, Hartmann A, Diep LM, Dahle DO, Reisaeter AV, Bjerve KS, Christensen JH, Schmidt EB, Svensson M. The association between marine n-3 polyunsaturated fatty acid levels and survival after renal transplantation. Clin J Am Soc Nephrol 2015;10(7):1246-56. doi: 10.2215/cjn.11931214. Kleber ME, Delgado GE, Lorkowski S, Marz W, von Schacky C. Omega-3 fatty acids and mortality in patients referred for coronary angiography. The Ludwigshafen Risk and Cardiovascular Health Study. Atherosclerosis 2016. doi: 10.1016/j.atherosclerosis.2016.06.049.

RI PT

97 98 99 100 101 102 103 104

AC C

EP

TE D

M AN U

SC

105

8