In general, diets high in red meat have been associated with increased CVD risk whereas higher intakes of fish and poultry have been associated with lower CVD risk (Hu et al., 2000; Fung et al., 2001; He et al., 2004a, 2004b). A report on 44 875 men aged 40–75 from the Health Professionals Follow-up Study found that a ‘prudent’ dietary pattern which included higher intakes of fruit, vegetables, whole grain, fish and poultry, was associated with a lower risk of CHD, whereas a ‘Western’ dietary pattern characterized by red meat, processed meat, high fat dairy products, refined grains, and sweets was associated with higher CHD risk (Hu et al., 2000). Similar associations of these dietary patterns with risk of CHD and stroke were reported by the NHS (Fung et al., 2004a, 2004b).
Much of the data on meat consumption and cardiovascular diseases come from longitudinal studies of California Seventh-Day Adventists (Snowdon et al., 1984; Snowdon, 1988; Fraser, 1999), a conservative religious group whose members are discouraged from consuming meat, fish, coffee, and eggs, and prohibited from using alcohol and tobacco. About half of all Adventists are lacto-ovovegetarians (i.e., exclude meat but not dairy and eggs from their diets), and there is considerable variation in past and current levels of meat consumption among vegetarians and nonvegetarian Adventists, respectively (Snowdon et al., 1984). Snowdon and colleagues reported the relationship of meat consumption and 21-year risk of ischemic heart disease (IHD) mortality among 25 153 men and women (baseline ages 45–84) from the Adventists Mortality Study, whose dietary habits were assessed by a questionnaire at baseline in 1960 (Snowdon et al., 1984; Snowdon, 1988). Usual meat consumption (specifically meat and poultry) was assessed by a single question, that is, ‘How many days per week do you eat meat?’ Fish consumption was assessed separately. Vegetarians were defined as individuals with meat/poultry intake less than 1 day a week. Because red meat and poultry intake was assessed together by a single question, differences in their association with risk could not be assessed. Meat consumption was positively associated with fatal IHD in both men and women. Furthermore, the observed higher age-adjusted risks of fatal IHD for nonvegetarian men and women were similar among those with versus without a history of heart disease at baseline, that is, 43% and 35% versus 49% and 37%, respectively. In subsequent analyses combining individuals with and without baseline heart disease and stratified by level of meat consumption (6 or more, 3–5, 1–2, and less than 1 day/week, i.e., vegetarians/reference group), meat intake was significantly and positively associated with IHD in all age groups except in the oldest group for men and in the two older age groups (ages 65–74 and 75–84) for women. Positive associations between meat consumption and IHD were stronger in men than in women, and, overall, strongest in younger (ages 45–54) than older men. In age-adjusted analyses, men who ate meat 6 or more days a week had almost a threefold higher risk of fatal IHD compared with those who ate meat less than 1 day a week; for women, risk (computed for the 45–64 year age group since number of deaths were small in the 45–54 age group) was nonsignificantly higher by 26%. In multivariate analyses adjusted for age, marital status, cigarette smoking history, obesity, and frequency of egg, cheese, milk, and coffee consumption, findings were similar. Daily consumption of meat was associated with 70% higher risk of IHD for men and 37% higher risk for women compared with no meat intake (p values were less than 0.001 and 0.02, respectively) (Snowdon et al., 1984; Snowdon, 1988). Meat consumption had no clear association with stroke among either men or women from the Adventists Mortality Study (Snowdon, 1988).
These findings were supported by results from the more recent Adventists Health Study (1976–1998) cohort of 34 192 California Seventh-Day Adventists aged 25 years and older, in whom baseline dietary assessment was conducted by a food-frequency questionnaire (including questions regarding 51 different foods or food groups). In multivariate analyses adjusted for age, smoking, BMI, exercise, hypertension, and intake of several foods, beef consumption was significantly associated with fatal IHD in men; risk of fatal IHD was about twofold higher in men who ate beef less than three times a week and exceeded twofold (p < 0.0001) in men who ate beef three or more times a week than in vegetarians. However, no associations of beef consumption and fatal IHD were observed among women. Furthermore, beef consumption was not significantly associated with the incidence of nonfatal myocardial infarction (MI) (Fraser, 1999). In the Oxford Vegetarian Study, 6115 non-meat eaters (mean age 38.7 þ 16.6) and 5015 meat eaters (mean age 39.3 þ 15.4) were recruited between 1980 and 1984 through the Vegetarian Society of the U.K. and the news media. Nonvegetarians were nominated by vegetarians from among their family and friends. Diet was assessed by a simple food frequency questionnaire. During 12 years of follow-up, non-meat eaters had significantly lower standardized mortality ratios for IHD than meat eaters; compared with the general population, mortality was 49% lower for meat eaters and 72% lower (p < 0.01) for non-meat eaters. The unadjusted risk of IHD death for non-meat eaters was 45% lower than that for meat eaters. However, with adjustment for smoking, BMI, and social class, the association weakened and became nonsignificant (Thorogood et al., 1994; Applyby et al., 1999). In a cohort of 4336 men and 6435 women (43% vegetarians) followed on average for 16.8 years, a vegetarian diet was associated with 15% lower mortality from IHD, but these findings were not significant (Key et al., 1996). A report, which combined data for 76 172 men and women from five prospective cohort studies, compared mortality rates between vegetarians (persons who did not eat any meat or fish) and nonvegetarians with other similar lifestyles. IHD mortality rate was 24% lower in vegetarians than in nonvegetarians after a mean follow-up of 10.6 years. Compared with persons who ate meat at least once a week, IHD mortality was 20% lower in persons who ate meat occasionally but less than once a week, 34% lower in persons who ate fish but not meat, 34% lower in lactoovovegetarians, and 26% lower in vegans (Key et al., 1999).
Menotti and colleagues reported associations between food groups and a 25-year risk of CHD mortality among 12 763 middle-aged men from the Seven Countries Study. Animal food groups were directly correlated and fish and vegetables were inversely correlated with CHD mortality (Menotti et al., 1999). Among women who participated in the NHS, consumption of red meat was associated with significantly higher risk of CHD by 43% per each 1serving/day, whereas consumption of poultry and fish was associated with 49% lower risk of CHD after adjustment for age. These associations were attenuated and became nonsignificant after multivariate adjustment. The ratio of red meat to poultry and fish was more strongly associated with CHD risk and remained significant after multivariate adjustment; CHD mortality risk was 32% higher for the highest compared with the lowest quintile of intake ( p–trend, 0.001) (Hu et al., 1999). In a report from the Iowa Women’s Health Study on 29 017 postmenopausal women, Kelemen and colleagues (2005) estimated a 15-year risk of CHD mortality from a simulated isoenergetic substitution of dietary protein for carbohydrates and of vegetable for animal protein (with adjustment for carbohydrates). Among women in the highest quintile of intake, CHD mortality risk was significantly lower by 30% from the substitution of vegetable for animal protein compared with women with the lowest quintile of intake ( p–trend, 0.02). No association with risk of CHD mortality was seen when animal protein was substituted for carbohydrates. However, CHD mortality was significantly associated with red meats; risk of CHD mortality was 44% higher for the highest versus lowest quintile of intake when red meat was substituted in place of number of servings of carbohydrate-rich foods ( p trend, 0.02). No association with poultry or fish was observed.
Sauvaget and colleagues (2003) examined the associations of animal product intake with risk of stroke mortality in a cohort of 37 130 men and women from Japan, who responded to a mailed survey, including a 22-item food frequency questionnaire, and who were followed for 16 years. Consumption of beef and pork or of pork products was not associated with higher risk of stroke mortality. A positive but nonsignificant association of chicken intake with stroke mortality risk was observed (i.e., 43% higher risk for almost daily consumption versus none). Among participants with the highest tertile of intake of animal products combined (beef and pork, chicken, pork products, milk, eggs, dairy products, fish, broiled fish), risk of all stroke mortality and risk of intracerebral hemorrhage mortality was significantly lower by 12% and 24%, respectively ( p–trend, 0.03), compared with persons with the lowest tertiles of intake. In a case-control study of lifestyle risk factors, stroke cases were obtained from a population-based register of acute cerebrovascular events in Perth, Western Australia. Consumption of meat more than four times a week was associated with higher risks of all strokes combined and first-ever strokes, in persons without prior history of any stroke or transient ischemic attack ( Jamrozik et al., 1994). However, when controls were followed prospectively for 4 years, consumption of meat more than four times a week was associated with lower risk of fatal vascular events (deaths from CHD, stroke, ruptured aortic aneurysm, peripheral vascular disease, or mesenteric thrombosis) by 38%, major vascular events (fatal vascular events, nonfatal MI, or nonfatal stroke) by 40%, and first-ever major vascular events by 44%, compared with less intake ( Jamrozik et al., 2000).
Fish Consumption And Cardiovascular Disease
Fish And CHD
While numerous studies have investigated the association of fish consumption with CHD mortality, there are little data from randomized clinical trials. Prospective epidemiologic studies have yielded inconsistent findings. Some studies showed an inverse association between fish intake and risk of CHD mortality (e.g., Kromhout et al., 1985; Norell et al., 1986; Kromhout et al., 1995; Oomen et al., 2000; Mozaffarian et al., 2003; Jarvinen et al., 2006), whereas others found no association (e.g., Curb and Reed, 1985; Vollset et al., 1985; Fraser et al., 1992; Salonen et al., 1995; Osler et al., 2003). Among 1822 men aged 40–55 and free of CVD at baseline from the Chicago Western Electric study, there was an inverse graded association between fish consumption and 30-year mortality from CVD and CHD, particularly nonsudden death from MI. In multivariate-adjusted analyses, risk of CHD mortality was lower by 38%, 16%, and 12% for men who consumed 35 g or more, 18–34 grams, and 1–17 grams of fish per day, respectively, compared with those who consumed none ( p trend, 0.04). Risk of death from MI (sudden and nonsudden) was lower by 44%, 24%, and 12%, respectively (p trend, 0.02). The observed inverse association of fish consumption and CHD mortality risk was accounted for by the relation of fish consumption to nonsudden death from MI; men who consumed 35 or more grams a day had a 67% lower risk (p trend, 0.007) compared with those who did not consume fish (Daviglus et al., 1997).
Among 20 551 male U.S. physicians ages 40–84, free of MI, CVD, and cancer at baseline, consumption of fish at least once a week was associated with a 52% lower 11-year risk of sudden cardiac death compared with less than once a month fish consumption (analyses were adjusted for potential confounders; p, 0.04). Lower risk of sudden death was seen at all levels of fish consumption, but the magnitude of risk reduction did not differ substantially at intake levels greater than one serving week, suggesting a possible threshold effect. Fish consumption and omega-3 fatty acids were not associated with lower risk of total MI, nonsudden cardiac death, or total CVD mortality (Albert et al., 1998).
Among 18 244 Chinese men aged 45–64 years followed for an average of 12 years, those who consumed at least one serving of fish and shellfish ( 50g) a week had a 44% lower risk of fatal MI compared with those with less intake. In analyses adjusted for age, total energy intake, and known CVD risk factors, men who consumed 200 or more grams (four or more servings) of fish/shellfish a week had a 59% lower risk of fatal MI compared with those who consumed less than 50 grams a week. Seafood intake was not associated with risk of death from stroke or IHD other than acute MI. In analyses examining fish and shellfish intake separately, the inverse associations with acute MI persisted and were significant for both fresh/ salted fish intake (70% of total seafood consumed) and shellfish intake (30% of total seafood consumed) (Yuan et al., 2001).
Among 41 578 middle-aged Japanese men and women, higher fish intake was significantly associated with lower 11-year CHD risk, primarily nonfatal cardiac events, compared with moderate fish intake of once a week (about 20 grams/day) after adjustment for CVD risk factors and selected dietary variables. Risks of total CHD and definite MI were lower by 37% and 56%, respectively, among persons in the highest quintile of fish intake (8 times/week, median intake 180 grams/day) compared with those in the lowest quintile (once/week, median intake 23 g/day). The inverse association was primarily seen for nonfatal CHD events (risk was significantly lower by 57%) but not for fatal CHD events. These findings suggest that higher levels of fish intake can even further reduce the risk of initial CHD events compared with moderate intake (Iso et al., 2006). Higher fish consumption was also significantly associated with lower 16-year risk of CHD morbidity and mortality among 84 688 U.S. women aged 34–59 years who were free of cancer and CVD at baseline. Compared with women who rarely ate fish (less than once a month), multivariate-adjusted risk of CHD (CHD deaths and nonfatal MI) was lower by 21%, 29%, 31%, and 34% for those who ate fish one to two times a month, once a week, two to four times a week, and five or more times a week, respectively ( p trend, 0.001). The inverse association was stronger for risk of CHD mortality (p trend, 0.01) than for nonfatal MI ( p trend, 0.10); women who ate fish five or more times a week had a significantly lower risk of CHD mortality by 45%, whereas their risk of nonfatal MI was nonsignificantly lower by 23% (Hu et al., 2002).
A review published in 1999, which quantified the fish intake–mortality relationship based primarily on only four cohort studies, reported that the inverse association between fish consumption and CHD mortality was evident among ‘high-risk’ but not ‘low-risk’ populations. The authors concluded that consumption of up to 40 to 60 grams of fish daily had a dose-dependent association with markedly lower CHD mortality for individuals who were high risk (i.e., had adverse risk factors or unhealthy lifestyle behaviors), but that there were no added benefits of fish consumption for persons at low risk for CHD and with a healthy lifestyle (Marckmann and Gronbaek, 1999). However, a recent meta-analysis of 13 cohorts from 11 independent cohort studies comprising 222 364 participants (3032 CHD deaths) with an average 11.8 years of follow-up found a consistent inverse association between fish consumption and CHD mortality rates. Compared with those who never consumed fish or ate fish less than once a month, individuals with a higher fish intake had lower CHD mortality: in pooled multivariate analyses, risks for CHD mortality were 11%, 15%, 23%, and 38% lower with fish intake of one to three times a month, once a week, two to four times a week, and five or more times a week, respectively, compared with no fish intake. Inverse associations were more evident among studies with a follow-up period of 12 years or longer. A dose–response relationship between fish consumption and CHD mortality was observed: that is, each 20 grams/ day higher intake of fish was associated with a 7% lower risk of CHD mortality ( p trend, 0.03) (He, 2004). Another meta-analysis of 19 observational studies (including 14 cohort and 5 case-control studies) found that risks of fatal CHD and total CHD were lower by 17% and 14%, respectively, for those who consumed any fish compared with little or none ( p < 0.005 for both) (Whelton et al., 2004).
The benefits of fish consumption on CHD mortality rates have also been demonstrated in diabetic populations. Among 5103 diabetic women from the NHS who were free of any history of heart disease or stroke at baseline, higher consumption of fish was associated with lower CHD and total mortality, even after adjustment for established cardiovascular risk factors. Compared with women who rarely ate fish (<1 serving/month), risk of CHD (adjusted for age, smoking, and other CHD risk factors) was 30%, 40%, 36%, and 64% lower with fish consumption one to three times a month, once a week, two to four times a week, and five or more times a week, respectively (p trend < 0.002). The inverse association persisted with adjustment for dietary factors related to CHD, including fiber, trans-fatty acids, polyunsaturated to saturated fats ratio, or fruits, vegetables, and red meat intake. Consumption of fish five or more times a week was associated with a significantly lower risk of fatal CHD by 59% and of nonfatal MI by 72% compared with intake of less than one serving a month. Additional adjustment for fruits, vegetables, and red meat did not affect the estimated risks (Hu et al., 2003). Moreover, fish consumption has been associated with significantly reduced progression of coronary atherosclerosis in diabetic women with CHD. Among 229 postmenopausal women with coronary atherosclerosis, changes in the mean minimum coronary artery diameter, the mean percentage of stenosis, and the development of new coronary lesions were evaluated by quantitative coronary angiography at baseline and after 3.2±0.6 years. Compared with lower fish intakes, consumption of two or more servings of fish or one or more servings of tuna or dark fish a week was associated with smaller increases in the percentage of stenosis (4.54±1.37% vs. -0.06±1.59% and 5.12± 1.48% vs. 0.35 ± 1.47%, respectively; p < 0.05 for both) in diabetic women after adjusting for age, CVD risk factors, and dietary intakes of fatty acids, cholesterol, fiber, and alcohol. No significant associations were found in nondiabetic women. Higher fish consumption was also associated with smaller declines in minimum coronary artery diameter and fewer new lesions (Erkkila et al., 2004).
Fish And Stroke
Several prospective epidemiologic studies have examined the association between fish consumption and the risk of stroke. However, the findings have been conflicting, with some cohort studies reporting an inverse association between fish intake and risk of stroke after adjustment for potential confounders (e.g., Keli et al., 1994; Iso et al., 2001; Mozaffarian et al., 2005), while others have reported no significant associations (e.g., Folsom and Demissie, 2004; Nakamura et al., 2005; Myint et al., 2006). In the Health Professional Follow-up Study (43 671 men aged 40–75 years and free of CVD at baseline), consumption of fish one to three times a month was associated with 43% lower risk of ischemic stroke compared with less frequent consumption. However, a higher frequency of fish intake was not associated with further risk reduction. No significant associations were found between fish intake and risk of hemorrhagic stroke or of total stroke (He et al., 2002). In the Zutphen Study, men who ate more than 20 grams of fish a day had about a 50% lower risk of total stroke compared with those who consumed less (Keli et al., 1994). A case-control study of lifestyle risk factors for stroke, involving cases from Perth, Western Australia, reported an inverse relation of fish consumption to first stroke; with risk significantly lower by 40% with consumption of fish more than twice a month compared with less consumption. In identical multivariate risk factor models for ischemic stroke and for intracerebral hemorrhage, a significant inverse relation with fish consumption was found for intracerebral hemorrhage (risk was significantly lower by 57%) but not for ischemic stroke (risk was nonsignificantly lower by 10%) ( Jamrozik et al., 1994). However, in a cohort of Japanese men and women, risks of mortality from total stroke and intracerebral hemorrhage were significantly lower by 15% and 30%, respectively, for the highest tertile of intake compared with the lowest (Sauvaget et al., 2003).