Circulating Docosahexaenoic Acid and Risk of All-Cause and Cause-Specific Mortality

Abstract

Objective:

To assess the associations of docosahexaenoic acid (DHA), a marine omega-3 fatty acid, with long-term all-cause mortality, cardiovascular (CV) mortality, and cancer mortality.

Patients and Methods:

We analyzed data from UK Biobank, which included 117,702 subjects with baseline plasma DHA levels and 12.7 years of follow-up between April 2007 and December 2021. Associations with risk for mortality endpoints were analyzed categorically by quintile of DHA plasma levels.

Results:

Comparing the lowest to highest quintiles of circulating levels of DHA, there was 21% lower risk of all-cause mortality (HR, 0.79; 95% CI, 0.74 to 0.85; P<.0001). In a secondary analysis, we merged the UK Biobank findings with those from a recent FORCE (Fatty Acid and Outcome Research Consortium) meta-analysis that included 17 prospective cohort studies and 42,702 individuals examining DHA and mortality associations. The cumulative sample population included 160,404 individuals and 24,342 deaths during a median of 14 years of follow-up. After multivariable adjustment for relevant risk factors comparing the lowest to the highest quintiles of DHA, there was 17% lower risk of all-cause mortality (95% CI, 0.79 to 0.87; P<.0001), 21% lower risk for CV disease mortality (95% CI, 0.73 to 0.87; P<.001), 17% lower risk for cancer mortality (95% CI, 0.77 to 0.89; P<.0001), and 15% lower risk for all other mortality (95% CI, 0.79 to 0.91; P<.001).

Conclusion:

Higher DHA levels were associated with significant risk reductions in all-cause mortality, as well as reduced risks for deaths due to CV disease, cancer, and all other causes. The findings strengthen the hypothesis that DHA, a marine-sourced omega-3, may support CV health and lifespan.

Acomprehensive meta-analysis of 38 randomized controlled trials comprising 149,051 participants recently showed that treatment with marine omega-3 fatty acids including docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) reduced risk for cardiovascular disease (CVD) mortality and fatal myocardial infarction.¹ However, the effects of omega-3 on non-CVD mortality rates and all-cause mortality remain uncertain. In the Cardiovascular Health Study, higher circulating levels of omega-3 were linked with “healthier aging” — defined as surviving past 65 years of age, free of chronic diseases, and maintaining good functional status.² A recent pooled analysis of data from 17 prospective cohort studies that included 42,702 individuals reported that the risk of death from all causes after multivariable adjustment was 15% to 18% lower (P<.003) in the highest compared with the lowest quintiles of circulating DHA and EPA levels.³

Blood levels of DHA and EPA are largely determined by chronic ingestion of foods and/or supplements containing marine omega-3s, although synthesis from dietary alpha linolenic acid makes a small contribution.⁴ If higher blood levels of DHA are confirmed to be associated with lower risk for all-cause and cause-specific mortality, it would provide support the American Heart Association’s recommendation to increase intake of marine omega-3s, which consistently raises DHA and EPA blood levels.⁵

This study aims to determine the relationship between in vivo DHA levels and risk for all-cause and cause-specific mortality in the United Kingdom Biobank (UKBB). Blood levels of DHA but not EPA were available in ~25% of the individuals enrolled in the UKBB.

The secondary purpose was to update a previous meta-analysis from the Fatty Acid and Outcome Research Consortium (FORCE) exploring the relationship between omega-3 fatty acids and total and cause-specific mortality and merge that data with new UKBB data. That study³ was a pooled, harmonized, de novo individual-level analysis of data from 17 prospective cohort studies.

PATIENTS AND METHODS

UKBB Sample

The UKBB is a large-scale biomedical database and research resource that contains in-depth genetic and health information from approximately 500,000 thousand participants from England, Scotland, and Wales.⁶ The cohort was recruited between April 2007 and December 2010 and were followed from April 2007 to December 2021. Measurement of DHA (DHA%) was performed in plasma as a percent of total fatty acids by nuclear magnetic resonance.⁷ Within the complete cohort of 502,639 individuals who were 40 to 69 years of age, a random subset of 117,702 adults 18 years of age or older (approximately 25%) were selected for analysis of plasma DHA% and did not die within 1 year of recruitment. Ethical approval for the UKBB is specified in UK Biobank Supplemental Material, (the full UKBB protocol is available online).^8,9

Statistical Methods

The primary outcome was mortality, with cause-specific mortality categorized as cardiovascular (CV)-related, cancer-related, or other. Sample characteristics were summarized using standard descriptive metrics (eg, means, SDs, percent, etc). Cox proportional hazards models were used to predict risk for death during the follow-up period adjusting for the 16 demographic and medical history covariates listed in Supplemental Table 1 (available online at http://www.mayoclinicproceedings.org). All variables were available in the UKBB primary release dataset. Separate models were fit for all-cause mortality, CVD mortality, cancer mortality, and other mortality. The relationships between DHA status and mortality outcomes were examined by quintile of plasma DHA%. A P value for linear trend across the five quintiles was also estimated.

A previously published meta-analysis of 17 cohorts³ was updated with the present findings from the UKBB using inverse-variance-weighted meta-analysis. All analyses were performed using R, with meta-analysis using the metafor package.¹⁰ A two-tailed P less than .05 was statistically significant.

RESULTS

UKBB Sample Characteristics

The 117,702 UKBB subjects experienced 8622 deaths during the follow-up period (median, 12.7 years; maximum, 14.6 years), with 1831 CVD-related deaths, 4183 cancer-related deaths, and 2609 deaths due to other causes. Supplemental Table 1 provides an overview of the sample at baseline. On average, the UKBB participants were 57 years of age, nearly evenly split by biological sex (54% female), primarily Caucasian (94%) and were overweight (mean body mass index = 27 kg/m²). The sample was relatively healthy at baseline with prevalence of hypertension, diabetes, and dyslipidemia each less than 20%, and self-reported health levels of good or excellent for nearly three-quarters of the sample. The proportion of patients per quintile taking omega-3 supplements is outlined in Table 1.

TABLE 1.

Plasma DHA% by Quintile^a,b

Quintile	DHA% range	Median DHA%	Estimated median omega-3 index, %	% Reporting fish oil supplement use
Q1	<1.47	1.25	3.75	18.3
Q2	1.47–1.79	1.64	4.96	23.8
Q3	1.79–2.08	1.93	5.86	29.6
Q4	2.08–2.48	2.26	6.89	37.9
Q5	>2.48	2.85	8.72	47.2

^aDHA, docosahexaenoic acid; Q, quintile.

^bEstimated omega-3 index (ie, the sum of EPA and DHA in red blood cell membranes expressed as a percent of total membrane fatty acids¹¹) and fish oil use.

DHA and Mortality in UKBB

Levels of DHA were significantly and inversely associated with all-cause and cause-specific mortality. When comparing the highest DHA quintile (Q5) to the lowest (Q1), there was a 21% lower risk of all-cause, CVD, cancer, and other mortality (Figure 1; Supplemental Table 2, available online at http://www.mayoclinicproceedings.org).

FIGURE 1.

Associations of circulating docosahexaenoic acid levels by quintile with all-cause and cause-specific mortality in the UK Biobank. The y-axis is truncated at 50% risk reduction.

These findings in the UKBB were then added to those of a previous meta-analysis examining the same relationships that included data from 17 cohorts (Table 2, Figure 2).³ Here, comparing Q5 to Q1, risk for all-cause mortality was lower by 17%; for CVD death lower by 21%; for cancer death lower by 19%; and death from all other causes lower by 15% (all P<.0001). P values for were for linear trends across the quintiles were all less than .001.

TABLE 2.

Association Between Circulating DHA Levels and Risk for All-Cause and Cause-Specific Mortality^a,b

	All-cause mortality HR (95% CI)	CVD mortality HR (95% CI)	Cancer mortality HR (95% CI)	Other mortality HR (95% CI)
Q1	1.0	1.0	1.0	1.0
Q2	0.94 (0.91–0.98)c	0.95 (0.88–1.02)	0.93 (0.89–0.99)d	0.94 (0.88–1.00)
Q3	0.90 (0.87–0.94)e	0.84 (0.77–0.90)e	0.90 (0.85–0.95)f	0.93 (0.87–1.00)d
Q4	0.94 (0.92–0.97)f	0.90 (0.83–0.97)c	0.86 (0.81–0.91)e	0.87 (0.81–0.93)e
Q5	0.83 (0.79–0.86)e	0.79 (0.73–0.87)e	0.81 (0.77–0.86)e	0.85 (0.79–0.91)e
P for linear trend	<.0001	<.001	<.0001	<.001

^aCVD, cardiovascular disease; Q, quintile.

^bPrevious Meta-analysis Updated with UKBB Data

^cP<.01.

^dP<.05.

^eP<.0001.

^fP<.001.

FIGURE 2.

Associations of circulating docosahexaenoic acid levels with all-cause and cause-specific mortality in the updated 18-cohort meta-analysis.

DISCUSSION

This is the largest ever study to examine the relationship between DHA status and long-term mortality. In our analysis of the UKBB data, DHA blood levels were inversely associated with risk of death from all causes as well as deaths due to CVD, cancer, and other causes. Comparing individuals with the highest vs the lowest DHA levels, risk for death across all four outcomes were 21% lower.

These findings corroborate those of an earlier pooled analysis of 17 studies with participants followed for a median of 14 years. For example, the relative risk of all-cause mortality (comparing Q5 to Q1) was 15% lower in the original meta-analysis and 17% lower in the updated meta-analysis. For CVD mortality, the 21% lower risk remained the same. For cancer, risk was 14% lower in the original and 19% lower in updated meta-analysis, and for death from other causes, the values were 13% and 15% lower, respectively.

A few studies have evaluated self-reported dietary fish/seafood (or estimated omega-3) intake in relation to all-cause mortality, and they generally support our findings here.^12–15 Self-reported use of fish oil supplements was associated with a significantly lower risk for death from any cause in a previous UKBB study that included more than 427,000 individuals.¹² However, self-reported food and/or supplement intake is memory-dependent and potentially unreliable. A more accurate and objective measure of omega-3 intake is the blood level of DHA and/or EPA.⁴

We focused on DHA in this meta-analysis as it was the only specific omega-3 fatty acid level available in the UKBB because nuclear magnetic resonance technology was not able to reliably measure plasma EPA in this population.⁷ Blood levels of DHA but not EPA show strong statistically significant inverse associations with risk of Alzheimer disease.^16,17 On the other hand, EPA monotherapy has been shown to be effective in reducing risk for major adverse CV events.¹⁸ No similar trials of DHA monotherapy have been undertaken. Levels of EPA+DHA have been shown to be inversely associated with mortality³; however, whether EPA or DHA is more strongly associated with improved life expectancy remains uncertain.

Biological Plausibility

Omega-3 consumption has been shown to confer positive health effects on several key age-related biological processes, including improved CV and immune function, enhanced cognition, and augmented neuromuscular function.¹ Lower blood pressure (BP) levels and slower resting heart rates are associated with reduced risks for CVD and premature mortality.^19,20 Approximately 3 g of DHA+EPA per day will lower systolic BP by a mean of 4.5 mm Hg in people with hypertension and 2 mm Hg in people with normal BP at baseline¹⁹ and will decrease resting heart rate by approximately 5 beats/min.^20,21

The vagus nerve plays a fundamental role in CV and brain health and in optimal function of the immune and gastrointestinal systems.²² People with higher vagal tone have been shown to be at lower risk of all-cause and CVD mortality during long-term follow-up.²² Supplementing with omega-3 has been shown to augment vagal tone in a dose-dependent manner.^20,21,23

Accelerated telomere shortening is associated with age-related diseases such as CVD, cancer, and neurodegenerative disease.²⁴ Blood EPA+DHA levels strongly predict the rate of telomere shortening, and these fatty acids have been shown to lengthen telomeres in a human study.²⁵ The mechanisms by which these effects occur are not clear but may be related to omega-3-induced reductions in oxidative stress and chronic systemic inflammation.²⁴

Omega-3 intake may have a beneficial effect on mammalian target of rapamycin (mTOR) signaling in certain circumstances.²⁶ For example, omega-3 supplementation may modulate mTOR activity in skeletal muscle tissue and the central nervous system.²⁷ Because dysregulation of mTOR pathways have been implicated in several age-related illnesses including neurodegenerative disease, cancer, and diabetes, the ability of omega-3 to regulate mTOR activity might be playing a role in its favorable effects on survival.

Frailty and sarcopenia are strongly correlated with aging and mortality. Recent studies have reported that ingestion of EPA+DHA exerts a positive influence on skeletal muscle with clinically relevant gains in muscle size and strength.^28,29 Omega-3 intake also protects against age-related loss of muscle mass and prevents deterioration in mitochondrial respiration during periods of muscle disuse.²⁸

Therefore, the reduced risks for all-cause mortality, CVD mortality, and cancer mortality associated with high DHA blood levels noted in the current study may be due to the cumulative effects of omega-3 on BP, heart rate, vagal tone, telomeres, mTOR, skeletal muscle mass/strength, mitochondrial function, and myriad other factors.

Omega-3 Intake

The average adult American consumes less than one serving per week of fish/seafood. Accordingly, the mean intake of DHA+EPA in the United States is only approximately 100 mg/d²⁹ and the mean omega-3 index is approximately 5.4%.³⁰ The omega-3 index for the top quintile in the present study was approximately 8%, suggesting that level as a cardioprotective omega-3 target as proposed nearly 2 decades years ago. To increase an omega-3 index of 5.4% to 8% would require consumption of approximately 1000 mg/d of DHA+EPA, and to go from the lowest quintile in the current study, with an omega-3 index 3.5%, to 8% would require approximately 1600 mg/d of DHA+EPA.⁵ These consumption levels are readily achievable from dietary fish/seafood and/or omega-3 supplements.³¹

It takes approximately 6 ounces of salmon to provide 2 g of omega-3 fatty acids. Still, many individuals do not choose to eat fish and seafood regularly and, for these people, omega-3 supplements are an effective and inexpensive way to raise DHA and EPA blood levels into the protective range.⁵ Approximately one-half of the individuals in the highest quintile of DHA blood level in the UKBB study were taking an omega-3 supplement. Over-the-counter omega-3 supplements typically contain ~300 mg to 600 mg of DHA+EPA per capsule.

Strengths and Limitations

The use of DHA blood level (an objective omega-3 biomarker) rather than dietary questionnaires markedly increased accuracy of exposure assessment. Additionally, combining the data from 18 studies that included more than 160,000 individuals and 24,000 deaths during 14 years of follow-up also greatly improved the statistical power to evaluate all-cause mortality and the specific subtypes of mortality.

One limitation of the current study is the relative lack of diversity — most individuals were White, possibly reducing the generalizability to other races/ethnicities. Docosahexaenoic acid and covariates were measured a single time at baseline, and fluctuations over time could lead to misclassification potentially biasing the results in unpredictable directions. Even so, omega-3 blood levels have shown reasonably good reproducibility over time.³² The UKBB cohort is generally healthier than the UK population of the same age; however, the biomarker-disease relationships observed in the UKBB cohort have been deemed to be likely generalizable to the wider population.³³

Because this is an observational study, causation cannot be established. Although we adjusted for numerous relevant risk factors (eg, age, sex, occupation, education, physical activity, smoking, hyperlipidemia, hypertension, etc), residual confounding by other unmeasured variables always remains a possibility.

CONCLUSION

This analysis shows highly significant inverse associations between DHA and all mortality endpoints assessed. These findings support the American Heart Association Science Advisory recommendation to consume at least one to two seafood meals/week.

Abbreviations and Acronyms:

CV

cardiovascular

CVD

cardiovascular disease

DHA

docosahexaenoic acid

EPA

eicosapentaenoic acid

mTOR

mammalian target of rapamycin