Journal of Cardiovascular Translational Research

, Volume 5, Issue 1, pp 92–99

Plasma Omega-3 Polyunsaturated Fatty Acids and Survival in Patients with Chronic Heart Failure and Major Depressive Disorder

Authors

    • Department of MedicineDuke University Medical Center
    • Department of Psychiatry and Behavioral SciencesDuke University Medical Center
  • Harry Oken
    • Martek Biosciences Corporation
  • Mona Fiuzat
    • Department of MedicineDuke University Medical Center
  • Linda K. Shaw
    • Duke Clinical Research Institute
  • Carolyn Martsberger
    • Department of Psychiatry and Behavioral SciencesDuke University Medical Center
  • Maragatha Kuchibhatla
    • Center for AgingDuke University Medical Center
  • Rima Kaddurah-Daouk
    • Department of Psychiatry and Behavioral SciencesDuke University Medical Center
  • David C. Steffens
    • Department of Psychiatry and Behavioral SciencesDuke University Medical Center
  • Rebecca Baillie
    • Rosa & Co.
  • Michael Cuffe
    • Department of MedicineDuke University Medical Center
  • Ranga Krishnan
    • Department of Psychiatry and Behavioral SciencesDuke University Medical Center
    • Duke-NUS Graduate Medical School Singapore
  • Christopher O’Connor
    • Department of MedicineDuke University Medical Center
    • Martek Biosciences Corporation
  • for the SADHART-CHF Investigators
Article

DOI: 10.1007/s12265-011-9325-8

Cite this article as:
Jiang, W., Oken, H., Fiuzat, M. et al. J. of Cardiovasc. Trans. Res. (2012) 5: 92. doi:10.1007/s12265-011-9325-8

Abstract

The omega-3 fatty acid (FA) concentration is low in patients with coronary heart disease (CHD). Supplement of omega-3 FA improves cardiovascular outcomes in patients with CHD and heart failure (HF). However, plasma omega-3 FA and its role for prognosis in HF patients have not been examined previously. In this study, we explore the prognostic value of omega-3 polyunsaturated FA in HF patients with major depressive disorder (MDD). Plasma was obtained from HF patients with MDD who participated in the Sertraline Against Depression and Heart Disease in Chronic Heart Failure trial. FA methyl esters were analyzed by the method of a flame ionization detector. Weight percent is the unit of the omega compounds. The primary outcome was survival which was analyzed using Cox proportional hazards regression modeling. A total of 109 depressed HF patients had adequate volume for completion of the FA assays. Plasma total omega-3 (hazard ratio [HR] 0.65, 95% confidence interval [CI] 0.43–0.98) and EPA_0.1 unit (HR 0.73, 95% CI 0.56–0.96) were significantly associated with survival of patients with HF and co-morbid MDD. The results suggest that low plasma omega-3 FA is a significant factor for reduced survival in HF patients with MDD.

Keywords

Omega-3 fatty acidsHeart failureDepressionPrognosis

Introduction

Fish oil or omega-3 polyunsaturated fatty acid (FA) consumption has been shown to reduce the risk of major cardiovascular (CV) events and coronary heart disease (CHD) [15]. Supplement of docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), the main contents of fish oil, or omega-3 FA, have been found to reduce mortality [69] in CHD patients and patients with heart failure (HF) [10].

The role of plasma omega-3 FA in the prognosis of HF has not been examined. In a population of HF patients with major depressive disorder (MDD), we explored the relationship of plasma omega-3 FA with survival. We hypothesized that patients with lower levels of omega-3 FA would have increased risk of both CV events and death.

Methods

Study Population

This study consisted of a subset of patients who were recruited from the cardiac inpatient service at Duke University Medical Center for participation in the Sertraline Against Depression and Heart Disease in Chronic Heart Failure (SADHART-CHF) trial [11, 12] and agreed to participate in the biomarker study of the SADHART-CHF trial. All the study participants had clinical HF, defined by left ventricular ejection fraction (LVEF) ≤ 45% (within the previous 6 months), and New York Heart Association (NYHA) classes II–IV heart failure symptoms and met the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (DSM-IV) criteria for MDD.

After providing written, voluntary, informed consent, HF patients were screened for depressive symptoms using the Beck Depression Inventory Scale [13] and the Diagnostic Interview Schedule [11, 12]. A separate consent was obtained to provide a blood sample for the biomarker study. A patient's decision to participate in the biomarker study did not affect his or her participation in the SADHART-CHF trial. The protocol was reviewed and approved by the Duke University Institutional Review Board.

Vital Status Follow-Up

The SADHART-CHF trial and results have been described previously [11, 12]. Briefly, these HF patients with MDD were randomized 1:1 to sertraline 50 mg/day or matched placebo for a 12-week active treatment phase. The dose of the study medication was titrated up weekly to a maximum of 200 mg/day according to the participant's tolerability and depression response. During the 12-week intervention, all participants also received nurse-facilitated support. Regardless of completion of the 12-week active treatment phase, all participants were contacted at 6 months, 12 months, and annually thereafter to evaluate clinical events and vital status until the last enrolled participant completed a 6-month follow-up. The National Death Index was used to ascertain vital status of participants who could not be contacted, which allowed for a 100% completion of the survival endpoint. Events resulting in a procedure or hospitalization due to CV causes according to predefined criteria were considered CV events. All CV events were adjudicated by three cardiologists with an HF subspecialty who were blinded to treatment assignment [12].

Omega FA Assessment

Sample Collection

Blood samples were collected using a Vacutainer system (Becton, Dickinson and Company, NJ 07417) containing clot activator and gel serum separator and centrifuged at a g-force of 1,864 (3,000 rpm for 10 min at a fixed angle of 34° using a Sorvall Model RT 7 Plus Centrifuge). The resulting plasma was stored in Vacutainer tubes (Becton, Dickinson and Company, NJ 07417) containing ethylenediaminetetraacetic acid (EDTA) and frozen at −80°C. Once all plasma samples were collected, the tubes were coded in a blinded fashion and sent to the Martek Biosciences Clinical Laboratory for assay.

Plasma Phospholipid Extraction and FA Analysis

The methods for plasma phospholipid extraction and FA analysis have been described previously by Bailey-Hall et al. [14].

The FA profiles were expressed as a percent of the total μg of FA (weight percent). The plasma mean values of phospholipid α-linolenic acid (ALA) (C18:3n-3), DHA (C22:6n-3), and EPA (C20:5n-3) were calculated. The omega-3 proportion was computed from plasma phospholipid mean weight percent using the following equation: 100 × (C20:5n-3 + C22:5n-3 + C22:6n-3)/(C20:3n-6 + C20:4n-6 + C22:4n-6 + C22:5n-6 + C20:5n-3 + C22:5n3 + C22:6n-3). The total omega-3 includes the following fatty acids: 18:2n-6, 18:3n-6, 20:2n-6, 20:3n-6, 20:4n-6, 22:4n-6, and 22:5n-6. The total omega-6 includes the following fatty acids: 18:3n-6, 20:5n-6, 22:5n-6, and 22:6n-6. The ratio of omega-6 to omega-3 was calculated by dividing the total omega-6 by the total omega-3. The total omega-3 (ALA, DHA, EPA) total omega-6, and the ratio of omega-6 to omega-3 are the focus of the survival analyses.

Statistical Analysis

Baseline characteristics and omega-3 FA distributions were presented by a median and interquartile range for continuous variables and by a percentage for categorical variables. The primary endpoint of this analysis was mortality, and the secondary endpoint was a composite of CV events or death. Cox proportional hazards regression models were used to evaluate the relationship between omega FA with (a) time to death and (b) time to initial CV event or death. The association of age, sex, baseline LVEF, NYHA class, ischemic etiology of HF, baseline Hamilton Depression Rating Scale (HDRS) scores, treatment assignment (sertraline vs. placebo), or remission status with survival was assessed by Cox proportional hazards regression models as well. Because of the modest sample size and a low number of deaths, we elected to adjust for age, sex, and baseline ejection fraction (EF) in a multivariate analysis of the omega FA measures that were found to be significantly associated with survival. A P value < 0.05 was considered statistically significant. Statistical analyses were performed by statistical personnel within Duke University Medical Center, using statistical software, SAS, version 9.1 (SAS Incorporated, Cary, NC).

Results

Baseline Characteristics

During the time period the biomarker study was conducted, 264 depressed HF patients were enrolled in the SADHART-CHF trial; 135 (51.1%) of them provided plasma samples and 109 of them had enough volume stored in EDTA tubes that were available for FA assays. Compared to the SADHART-CHF participants who did not participate in the biomarker study [12], the 109 patients who participated in the biomarker study were younger (62 [55, 71] vs. 59 [55, 69] p = 0.013), were more likely to be male (57.5% vs. 69.7%, p = 0.013), had a lower rate of NYHA class III or IV (75.0% vs. 61.5%, p = 0.003), and had greater baseline HDRS scores (18 [14, 22] vs. 20 [15, 23], p = 0.039). The baseline characteristics and FA levels of the biomarker study population are summarized in Table 1.
Table 1

Baseline characteristics

Characteristics (N = 109)

 

Age (years)

60 (53, 68)

Male

76 (69.72%)

White

59 (54.13%)

LVEF (%)

30 (20,40)

NYHA class

 

 II

42(38.53%)

 III–IV

67(61.47%)

Ischemic etiology

69 (63.30%)

History of myocardial infarction (MI)

53(48.62%)

Baseline Beck Depression Inventory Scale

17(13, 22)

Total omega-3 (wt.%)

4.73% (3.93, 5.43)

Omega-3 proportion (wt.%)

19.10% (17.01,21.31)

ALA (wt.%)

0.23% (0.19, 0.29)

DHA (wt.%)

3.07% (2.37, 3.73)

EPA (wt.%)

0.45% (0.33, 0.59)

Total omega-6 (wt.%)

39.91% (38.21, 40.80)

Omega-6/omega-3 (wt.%)

8.37% (7.28, 10.34)

For categorical values, results are presented as N (%); for continuous values, results are presented as the median (25th and 75th percentile)

LVEF Left ventricular ejection fraction, NYHA New York Heart Association, ALA α-linolenic acid, DHA docosahexaenoic acid, EPA eicosapentaenoic acid

Omega Fatty Acids and Survival Status

Overall, the population had low omega-3, reflected by the omega-3 proportion (Table 1). The majority of the participants (79.8%) had their omega-3 proportion value <22%, which is considered a risk factor for sudden cardiac death and CHD [15]. There were only three patients (2.8%) that had an omega-3 proportion value equal or greater than 37%, which is an indicator for low risk of sudden cardiac death and CHD [15].

The median follow-up time was 496 days (IQR = 286–897 days). During this period, 23 (21.1%) patients from the cohort died. Compared with patients who remained alive, deceased patients had lower total omega-3 weight percent of FA (4.25% [3.66, 4.99] vs. 4.79% [3.99, 4.56] p = 0.009), DHA (2.69% [2.29, 3.40] vs. 3.11% [2.37, 3.87], p = 0.039), and EPA (0.33% [0.29, 0.47] vs. 0.47% [0.36, 0.59], p = 0.011). In contrast, ALA (0.25% [0.19, 0.29] vs. 0.23% [0.19, 0.28], p = 0.84), total omega-6 (39.56% [38.27, 40.60] vs. 40.0% [38.19, 40.81] p = 0.09), and the ratio of omega-6 to omega-3 (8.74% [7.63, 10.69] vs. 8.31% [7.08, 10.09] p = 0.54) did not differ between living and deceased patients. The Cox proportional univariate analysis demonstrated that total omega-3 (hazard ratio [HR] 0.65, 95% confidence interval (CI) 0.43–0.98) and EPA_0.1 unit (HR 0.73, 95% CI 0.56–0.96) were significantly associated with survival (Table 2). Multivariable Cox proportional models adjusting for age, sex, and baseline EF altered the significance of total omega-3 for HF survival (HR 0.68, 95% CI 0.452–1.026, p = 0.066). The significance of EPA for HF survival in the adjusted model remained (HR 0.76, 95% CI 0.573–0.972, p = 0.030).
Table 2

Survival analysis (univariate)

Variable name

Chi-square

HR

95% CI

P value

Total omega 3

5.3940

0.652

0.433

0.983

0.0202

Total omega 6

0.1642

0.964

0.812

1.145

0.6853

Omega 6/omega 3

3.3818

1.178

0.992

1.399

0.0659

DHA

3.5829

0.674

0.433

1.048

0.0584

EPA_0.1 unit

7.8671

0.733

0.560

0.960

0.0050

Age

0.7780

1.018

0.979

1.059

0.3778

Male

0.5436

0.719

0.305

1.699

0.4609

Treatment-Sertraline

3.1698

2.138

0.905

5.048

0.0750

NYHA

2.0158

1.527

0.854

2.731

0.1557

Caucasian

1.3899

0.610

0.267

1.394

0.2384

Married

0.1183

0.866

0.382

1.964

0.7309

Diabetes

1.3429

0.616

0.269

1.415

0.2465

Hypertension

0.7934

1.448

0.674

3.113

0.3731

Ischemic etiology

0.5674

1.257

0.709

2.228

0.4513

History of myocardial infarction

0.0586

0.932

0.523

1.661

0.8088

Baseline ejection fraction

2.5831

0.967

0.928

1.008

0.1080

Baseline Hamilton Depression Rating Scale

0.0291

1.007

0.932

1.088

0.8646

There were 65 (59.6%) patients from the cohort who experienced at least one fatal or non-fatal CV event with a total of 133 non-fatal CV events during the entire follow-up phase. These non-fatal CV events included unstable angina (5), MI (3), syncope (3), arrhythmia (12), stroke (2), HF exacerbation (70), and other CV events (38) that included renal failure or exacerbation of renal insufficiency due to HF, hypotension, or other complications due to HF intervention, etc. Seven of the 65 patients died without presentation of any non-fatal CV event. None of the omega-3 FAs were associated with event-free (time to death or first CV event) survival (Table 3).
Table 3

Event-free survival analysis

Variable name

Chi-square

HR

95% CI

P value

Total omega 3

1.4760

0.907

0.764

1.078

0.2244

Total omega 6

0.0130

0.994

0.896

1.103

0.9092

Omega 6/omega 3

0.7644

1.048

0.943

1.165

0.3820

DHA

2.0311

0.846

0.666

1.073

0.1541

EPA

0.8586

0.841

0.544

1.301

0.3541

Age

0.0163

1.002

0.978

1.025

0.8984

Male

0.1834

0.890

0.526

1.509

0.6685

Treatment-Sertraline

0.2235

0.889

0.546

1.448

0.6364

NYHA

0.3630

1.112

0.788

1.568

0.5469

Caucasian

3.3070

0.636

0.390

1.037

0.0690

Married

0.3527

1.159

0.713

1.886

0.5526

Diabetes

0.2287

0.889

0.548

1.443

0.6325

Hypertension

0.0367

1.057

0.603

1.852

0.8480

Coronary artery disease

0.0262

1.033

0.699

1.526

0.8713

History of myocardial infarction

0.2462

0.916

0.644

1.301

0.6198

Baseline ejection fraction

1.5081

0.985

0.962

1.009

0.2194

Baseline Hamilton Depression Rating Scale (HDRS)

0.0459

0.995

0.950

1.042

0.8303

Age, sex, baseline LVEF, NYHA class, ischemic etiology of HF, baseline HDRS scores, treatment assignment (sertraline vs. placebo), or remission status at the end of 12-week intervention were not significantly associated with either survival or CV-event-free survival (Tables 2 and 3).

Discussion

The major finding from this study is that lower total omega-3 and EPA concentrations were significantly associated with reduced survival for HF patients with MDD. The results suggest that every one unit lower in the total omega-3 or every 0.1 unit lower in EPA increases the risk of dying by 27–35% among HF patients with MDD over a median of 16 months. The association of total omega 3 and EPA with HF survival is independent of age, sex, and baseline EF. Furthermore, the majority of the study population has low omega-3.

Previous studies have suggested that blood and/or tissue proportions of specific FAs may stratify risk for a variety of diseases, especially CHD [1619]. In a review of 25 studies that examined the role of tissue FA and risk of CHD, Harris et al. demonstrated that DHA, among several other omega-3 FAs, was significantly reduced in patients who died of CHD [20]. Patients who experienced non-fatal CV events also had lower DHA or lower DHA + EPA combination, although it was not significant [20]. Because low DHA was the most common finding across all the studies reviewed, these authors [20] and others [21] concluded that DHA was more cardioprotective than EPA. There was only one study involving HF patients that tested the relationship between omega-3 FA and ventricular arrhythmia. This study found that 102 HF patients with implantable cardioverter-defibrillators (ICDs), compared with 25 control subjects, had higher red blood cell omega-3 FA, measured as omega-3 index, indicative of an association of higher omega-3 FA and increased risk for ventricular arrhythmia [22]. Our study focused on assessing the relationship of omega FAs with survival and did not measure heart rhythm irregularities during the follow-up. Furthermore, the omega-3 index was not calculated in our study, as the omega FAs were derived from plasma, not red blood cells.

Low concentrations of DHA and EPA have been reported in depressed patients without medical co-morbidity in peripheral tissues such as plasma [23], red blood cell phospholipids [24, 25], cholesterol esters [26], adipose tissue [27], and in the brain [28]. Similar to studies conducted in the cardiac population, the components of omega FAs and tissues tested for omega-3 levels are not consistent across studies. Conversely, fewer studies have examined the relationship of tissue omega concentration and depression. The major omega-3 FAs (ALA, EPA, and DHA) have been found to be low in cholesterol esters or red blood cells in depressed patients [23, 24, 26, 29]. Two larger sample studies [30, 31] demonstrated that depressed patients have lower DHA and EPA compared to normal controls.

There are several potential mechanisms [32] by which omega-3 FAs reduce the risk of CV diseases, such as stabilizing cell membranes, inhibiting production of thromboxane A2 and inflammatory cytokines [3335], increasing heart rate variability [36], etc. An echocardiographic study found omega-3 polyunsaturated FAs had a small but significant improvement in left ventricular ejection fraction in patients with HF [37]. Another recent mechanistic investigation demonstrated that 2 g omega-3 polyunsaturated FA of 11 months after a 5 g omega-3 polyunsaturated FA 1 month, compared to placebo, resulted in significant improvement in LVEF (increased by 10.4% and decreased by 5.0%, respectively); peak VO2 (increased by 6.2% and decreased by 4.5%, respectively); exercise duration (increased by 7.5% and decreased by 4.8%, respectively); and reduction of NYHA class (decreased from 1.88 ± 0.33 to 1.61 ± 0.49 and increased from 1.83 ± 0.38 to 2.14 ± 0.65, respectively) among 133 patients with non-ischemic HF. The intervention also resulted in lower re-hospitalization for HF vs. in placebo (6% vs. 30%, p = 0.0002) [38].

The understanding of the role of omega-3 FA in depression is evolving as well and evidence by far indicates that the vital roles of DHA and EPA in depression may be quite different. DHA is commonly believed to be the key element for mental health as it is an essential component of neuronal membranes and thus critical to the structure of the brain and nervous system, while EPA is believed to have direct interactions with cyclo-oxygenases, lipoxygenases, phospholipases, acylating systems, ion channels, mitochondria, and the peroxisome proliferator activated receptors. It is also thought to exert regulatory influences on gene expression [3941]. Nevertheless, more studies are needed to better understand the role of these FAs in CV disease and mental health.

Most patients in this study had low omega-3, indicated by the omega-3 proportion measurement, especially those who died during the follow-up. It is important to note that while DHA and EPA can be synthesized from ALA (Fig. 1) which comes from diet with plant oil, only a very small proportion of ALA can be converted to EPA [4244] and DHA [4245], due to the inefficiency of the process. HF patients with MDD, especially those who died during the follow-up, might have had lower fish oil intake or reduced activity of the desaturase and elongase enzymes that produce DHA and EPA. Decreased enzymatic activity could be due to diet and/or gene polymorphisms [46]. Desaturase gene polymorphisms were found to correlate with perinatal depression [47] and decreased gene expression levels correlated with suicide completion [48]. In addition, the aging brain has a decrease in desaturase and elongase activities as well as lower peroxisomal function effecting enzyme levels [4954].
https://static-content.springer.com/image/art%3A10.1007%2Fs12265-011-9325-8/MediaObjects/12265_2011_9325_Fig1_HTML.gif
Fig. 1

Biosynthetic pathway of omega-3 fatty acids

Clinical Implications

The Gruppo Italiano Di Studio Della Sopravvivenza Nella Insufficienza Cardiac-Heart Failure (GISSI-HF) trial has demonstrated that, compared with placebo, omega-3 FA supplementation of DHA + EPA (850 to 882 mg/capsule), improved survival (adjusted HR 0.91, 95% CI 0.83–1.00) and reduced CV re-hospitalizations (adjusted HR 0.92, 95% CI 0.85–1.00) over an average of 3.9 years of intervention among 6,975 HF patients [10]. The findings of our present study further support the importance of omega-3 FA supplements for HF patients, as lower omega-3 FA is associated with shortened survival of HF patients. A large number of studies have examined whether an omega-3 supplement may improve depression; however, the results are not as consistent [39, 55]. Results of meta-analyses indicate that omega-3 FA supplement is beneficial for patients who have a formal diagnosis of depression [56]. Variation of dosing and proportion of omega-3 FA supplement have been speculated to be responsible, at least partially, for the inconsistent findings of benefits in depression and many other conditions [39, 40, 55]. While Lovaza (similar EPA to DHA concentrations) failed to demonstrate antidepressant effects [57], an almost pure EPA supplement demonstrated antidepressant efficacy [58]. EPA has been argued to be the main component of omega-3 to alleviate depression [39, 40], but no definitive conclusions have been drawn [59]. Guiding omega-3 FA supplement therapy by using DHA and EPA concentrations may improve the outcome and reduce the heterogeneity of these trials.

Limitations

The study is limited by the origin of the sample to test if the same relationship of omega-3 and HF survival exists in HF patients without depression, and by the small sample size to fully evaluate confounders or mediators with the negative association of low omega-3 and high mortality. Recruitment for the biomarker study from depressed HF patients was challenging, reflected by the fact that only about half of the parent trial participants were willing to give blood. The study population was not large enough to fully assess the prognostic role of omega FAs for event-free survival. Further study is needed to examine the associations of omega-3 FA with occurrence of CV events or event-free survival in HF patients.

In conclusion, our study suggests that omega-3 FA has a significant association with survival in HF patients with MDD. Whether a supplement of omega-3 FA may improve depression as well as survival and other CV outcomes in these patients needs to be studied. The interplay between omega-3 and conventional risk factors as well as other biomarkers in patients with HF also warrants future study.

Clinical Relevance

Results of this study suggest omega-3 FA plays a significant role in the survival of patients with HF and MDD. Studies testing whether omega-3 supplement may improve depression and survival in this population are in demand.

Acknowledgements

The authors would like to thank Eileen Bailey-Hall of Martek Biosciences Corporation, Columbia, MD and the Martek Biosciences Corporation for the significant contribution in regards to processing plasma samples and the FA analysis. We would also like to thank Dr. Steve Rozen of the Duke-NUS Graduate Medical School Singapore in Singapore for his contribution of data analysis and interpretation.

Funding Sources

The study was funded by the SADHART-CHF study (R01-MH063211) and R21 MH076178 by the National Institute of Mental Health (NIMH) Bethesda, Maryland.

Financial Disclosures

• Dr. Jiang received salary support through the NIMH research grant, and a minority supplement award for the biomarker study affiliated to the SADHART-CHF grant.

• Dr. O'Connor has received salary support through the NIMH research grant and is a consultant for Martek Biosciences Corporation.

• Dr. Kuchibhatla received salary support through the NIMH research grant.

• Dr. Cuffe has received salary support through the NIMH research grant.

• Dr. Krishnan has received salary support through the NIMH research grant.

• Dr. Martsberger has received salary support from the NIMH research grant.

• Linda Shaw: nothing to disclose for this study.

• Dr. Fiuzat: nothing to disclose for this study.

• Dr. Oken is a consultant to the Martek Biosciences Corporation.

• Dr. Kaddurah-Daouk: nothing to disclose for this study.

• Dr. Steffens has received salary support from NIMH for the metabolomic study.

• Dr. Baillie: nothing to disclose for this study.

All authors had access to the data. The preliminary result of the omega fatty acid study was presented at the 2010 American College of Cardiology Annual Meeting in March 2010 in Atlanta, GA.

Copyright information

© Springer Science+Business Media, LLC 2011