Omega-3 Fatty Acids and Coronary Heart Disease: A Very Fishy Story

The new American Heart Association (AHA) Science Advisory, “Seafood long chain n-3 PUFAs and cardiovascular disease”¹ recommends one to two seafood meals per week to reduce risk of heart failure (HF), CHD, ischemic stroke, and cardiac death, especially when seafood replaces less healthy foods. It is unclear why one to two servings were selected. Twenty years ago, Connor² recommended two to three servings per week. The successful DASH diet recommended .5 servings per day or about three to four servings per week.³

The critical questions are whether this is the proper dose? Should recommended dose vary by host genetic susceptibility? Will this dose raise n-3 PUFAs levels in blood similar to those of low risk Japanese population? What is the effect of adding fish oil in a typical high fat, higher saturated fat US diet as compared to countries with high intake of n-3 PUFAs but also lower intake of saturated fat?

The authors of this statement and the recent previous reports of 2002 and 2016^4,5 were influenced by the results from clinical trials of n-3 PUFAs and supplements, especially in the US as compared to the ecological evidence based on much lower CHD rates in Japan with five or more times higher intake of n-3 PUFAs6. CVD rates are also low in other populations consuming higher amounts of n-3 PUFAs.

An important question is why is there such a discrepancy between both strong ecological inverse associations of high population dietary consumption of n-3 PUFAs and CHD incidence and mortality and the results of clinical trials.^4,6-11 Numerous, but not all, epidemiological studies have supported the benefit of fish consumption to reduce CHD/CVD death, except for fried fish.^2,12-18 Blood and tissue levels of n-3 PUFAs^19-25 are linked to lower risk of coronary artery disease (CAD). Clinical and laboratory feeding experiments have also found consistent positive benefits of n-3 PUFA consumption in relationship to inflammation, cardiac arrhythmias, left ventricular function, thrombosis, and plaque stability.^26-31 Pathology studies have also noted a low prevalence of atherosclerotic plaque associated with higher n-3 PUFAs.^32,33

The difference between the results of clinical trials and the ecological, epidemiological, and laboratory studies, on the benefits of n-3 PUFAs is far more than of academic interest having very substantial public health implications. The Seven Countries Study³⁴ as well as previous studies of populations in Greenland that consume large amounts of fish oils documented the low mortality from CAD. The Japanese consume about 1 g/day of n-3 PUFAs as compared to 100-200 mg/day in the US.^22,35 Plasma phospholipid-measured n-3 PUFA levels in the blood are much higher in the Japanese than in the US.^22,36 The key public health question is whether the recommended intake of seafood in the US should be to reach the same levels of n-3 PUFAs in blood as in Japan, about 9% of total fatty acids in blood versus 4% in the US. The addition of two servings of seafood per week might increase consumption of n-3 PUFAs by about 500 mg.²⁰

The most recent (2013) CHD age-adjusted death rates for Japanese men age 35-74 was 44/100,000 as compared to 128/100,000 for US men and for women, 12 versus 49/100,000, an over 4-fold difference.¹⁰ Endemic CHD in the US might practically disappear if consuming one g/day of fish oil, as is done in Japan, resulted in death rates in the US similar to Japan. There would be a substantial increase in life expectancy and likely decrease in health care utilization and costs for treatment of CVD. The low rates in Japan are not primarily due to genetic differences because migration of Japanese to the US or Brazil, for example, resulted in increase in CHD rates within a single generation associated with decline in n-3 PUFA intake.³⁷

Studies of post-World War II birth cohorts in Japan as compared to the US have documented higher blood pressure, cigarette smoking, and similar cholesterol levels in Japanese and US white men but much lower CHD rates and extent of coronary atherosclerotic disease, as measured by coronary computed tomography (CT). The slopes of the risk of CHD by risk factor are similar but at any level of CHD incidence was much lower in Japan.³⁸

Studies have consistently shown very low rates of CHD in southern Europe and in France^7,10,39 attributed to differences in diet, including higher intake of n-3 PUFAs, alpha linolenic acid, lower intake of saturated fat and higher fish intake and higher intake of specific monounsaturated fatty acids, i.e. olive oil, nuts, Mediterranean diet. ^40-45 The evidence strongly suggests that there is a protective effect in the diet that is associated with lower risk of CHD at similar levels of risk factors such as low-density lipoprotein cholesterol (LDL-C), Apolipoprotein-B, triglycerides, obesity. This contrasts with lower CHD rates among vegans due to lower intake of saturated fat and lower blood LDL-C levels.⁴⁶ Furthermore, the lower prevalence of atherosclerotic plaque and coronary artery calcium by CT in Japan versus the US would be consistent with an effect on development of atherosclerotic plaques rather than later effects primarily on thrombosis, fibrinolysis. However, the specific benefits, if any, of n-3 PUFAs is still unsubstantiated in clinical trials.

Nutritional epidemiology is an example of a common source epidemic similar to a population consuming a common food contaminated with a bacteria leading to cases of gastrointestinal disease.⁴⁷ The determinants of the epidemic include first, identifying the individuals who consume the specific common source; second, the dose of the agent; third, the likelihood of disease given the exposure, e.g. virulence; fourth, measures of host susceptibility, including genetics and other factors that either increase or decrease the likelihood of both disease and severity of disease given exposure; and fifth, the time between the exposure and the disease, the incubation period. In common source epidemics, it is often feasible to identify the specific common agent but very difficult to identify the differences in dose of exposure for those who did and did not develop the disease. This is especially true when many of the individuals may have subclinical disease, e.g. subclinical atherosclerosis.

Further investigation of n-3 PUFAs, therefore, as a common source exposure may improve our understanding of relationship to CHD:

1. The first issue is dose of exposure to the agent. The levels of triglycerides in Japan are similar to those in the US, even in the younger age groups in spite of the fact that the Japanese are much thinner than the US population. Substantial effects of n-3 PUFAs on triglyceride levels usually require doses higher than in the general population as noted, perhaps 2-5 g/day. Most clinical trials in the US except those treating hypertriglyceridemia include a maximum of only about 1 g/day, the average in Japan, or even lower doses. The dose in many of the trials may be too low to modify risk of CHD, especially in secondary prevention. Therefore, much higher dose of n-3 PUFAs clinical trials may be required.⁴⁸ There is little evidence that there are differences in blood fatty acid changes after consuming similar doses of n-3 PUFAs from fatty fish versus capsules containing fish oils.²⁰
2. Are there other dietary sources that impact the association of n-3 PUFAs with CHD risk? Countries with low incidence of CHD, such as Japan as well as in southern Europe, also generally consume lower amounts of animal products, especially beef, and have lower saturated fat intake.^49,50 The virulence, e.g. the development of atherosclerosis and clinical CVD, may be a function of both the amount of consumption of saturated fat and other nutrients in animal products as well as the amount of n-3 PUFAs and other poly- and monounsaturated fatty acids in the diet.^51,52 The amount of n-3 PUFAs in diet or supplements to reduce atherosclerosis and CVD may have to be much higher if the saturated fat intake of the population is also very high, e.g. higher than it is in Japan. For example, feeding studies many years ago from the Cleveland Clinic and others showed that the amount of polyunsaturated fatty acids necessary to reduce LDL-C was a function of the amount of saturated fat and polyunsaturated fat at the current levels of 10-15% of saturated fat, polyunsaturated fat had to increase to 12-15% or more to have a major beneficial effect on the LDL-C levels and probably the reduction of CHD, much higher than were used in some of the trials and much higher than the intake of n-3 PUFAs in the US population.⁵³ It is of great relevance possibly that two major trials of Mediterranean diet and n-3 PUFAs (Japan EPA Lipid lowering Interventions Study [JELIS]) were carried out in low CHD risk populations, e.g. Spain and Japan. Neither trial, however, was double blinded and limits the interpretation of results.^44,54 The JELIS trial used higher doses of n-3 PUFAs in a population with higher intake of n-3 PUFAs and the PREDIMED trial (Prevención con Dieta Mediterránea) provided both olive oil and nuts to intervention group. Would either of the trials had the same benefit effect if tested in a population consuming a typical higher saturated fat and protein, low polyunsaturated fat diet, e.g. in the US? Would increases in alpha linolenic acid n-3 PUFA from vegetable sources reduce CHD incidence if just added to a usual US diet?55,56 There is, for example, suggestion that benefit of alpha linolenic acid (vegetable source) may be a function of levels of arachidonic acid (animal source polyunsaturated fat).⁵⁷ Future clinical trials and feeding studies in the US should evaluate the effects of very high n-3 PUFA intake and lower saturated fat intake, e.g. similar to Japan, France, or higher monounsaturated fat, e.g. southern Europe. The key limiting variable may be the amount of saturated fat in the diet.
3. Is the incubation period for n-3 PUFAs exposure too short in most clinical trials? The Japanese and other populations with low CHD rates consume potentially beneficial high n-3 PUFA diets, etc. from a very early age. The prevention of CVD may depend on the effects on early stages of atherosclerotic development or the rate of progression of atherosclerosis. Feeding n-3 PUFAs in clinical trials to reduce CHD events among older participants with extensive atherosclerosis for 1, 2, or even 5 years may miss the need for long term interventions beginning possibly at earlier ages because of the long incubation period of atherosclerosis to clinical CAD.
4. Host susceptibility may be an important determinant of benefits of n-3 PUFAs. The nutrients in the western diet, e.g. related to the high consumption of animal protein and saturated fat that may affect the gastrointestinal biology and metabolism and contribute to the high rates of CHD seen in the US and in many other countries. Differences in metabolism of nutrients may be affected both by differences in GI microflora among countries and interaction of genetic and unique environmental exposures, e.g. epigenetic effects. Further feeding studies could focus on n-3 PUFAs, phospholipid metabolism and relationship to atherosclerosis.^58-61

It is also possible that higher doses of n-3 PUFAs have a primary anti-inflammatory role in slowing the progression of atherosclerosis or possibly on thrombosis and fibrinolysis. Better short term feeding studies can now evaluate the effects of n-3 PUFAs on new measures of both adaptive and innate immunity and atherosclerosis.^62-65 Similarly, ecological studies could evaluate the relationship of incident morbidity and mortality of other diseases related to inflammation, e.g. rheumatoid arthritis, other “autoimmune diseases,” COPD, etc. in relation to consumption of n-3 PUFAs.

The current design of clinical trials may have provided false-negative interpretation of the relationship between n-3 PUFAs and reduction of CHD incidence or mortality. On the other hand, it is possible that both the ecological and the clinical epidemiological studies may be incorrect, that the substantially lower CHD rate in Japan may be a function of their lower body weight, continued lower saturated fat intake, etc. or to other factors, such as the high intake of soy flavonoids. We also may be still observing a cohort effect and that the prior low levels of cholesterol, cigarette smoking, and saturated fat intake are still impacting on the population even these many years later.^66-69 However, studies of birth cohorts in Japan born after World War II who had substantial changes in their risk factors early in life to levels similar to the US still consistently show both higher intake of n-3 PUFAs and substantially lower extent of atherosclerosis and CHD mortality.^38,70-72

In summary, CHD is an example of a common source epidemic driven primarily by nutrients that account for the continued marked geographic variations in CHD in Japan, France, Israel, southern Europe. Genetic, host susceptibility, and other risk factors, i.e. smoking, elevated BP, physical activity, and obesity, obviously contribute to risk of CHD. Nutrients and their effect on lipoprotein metabolism, inflammation, and thrombosis is the driving force for this common source epidemic. Further nutrition research is important that could improve dietary guidelines and recommendations for seafood consumption in the US.

Better and longer feeding experiments with measures of intermediate endpoints should be considered. The choice of intermediate endpoints is problematic, lipoproteins, inflammation, thrombogenesis, vascular imaging. The availability of new technologies to measure intermediate endpoints may enhance nutrition feeding studies. The ideal clinical trial mimicking the experience in Japan would have to enroll young adults and follow them for at least 20-30 years. Such a trial is not feasible. Current trials may improve our understanding of effects of high doses of n-3 PUFAs in the short term, especially among individuals with atherosclerosis and/or CHD but will not resolve the issue of dose for public health prevention of CHD.

The increase in seafood or other sources of n-3 PUFAs may not be feasible at the population level without input from the food industry. Modern genetic technologies could result in foods with higher n-3 PUFAs content and reduction of saturated fat.

The marked geographic variations in CHD incidence and mortality, especially in Japan, France, and in southern Europe as compared to the US and other countries still persist. Reduction of CHD rates in the US to the low incidence and mortality countries like Japan would have, by far, the biggest impact on the CHD epidemic in the US.^73-75 A common source epidemic cannot be controlled by individual-based interventions alone. It is perhaps unfortunate that after 50+ years of research, the best we can do is recommend one to two servings of seafood per week. A further substantial decline in incidence and mortality due to CVD is very likely to depend on modification of the common source diet.⁴⁵