Fish and Omega-3 Fatty Acid Intake and Risk of Coronary Heart Disease in Women

Context Higher consumption of fish and omega-3 fatty acids has been associated with a lower risk of coronary heart disease (CHD) in men, but limited data are available regarding women.

Objective To examine the association between fish and long-chain omega-3 fatty acid consumption and risk of CHD in women.

Design, Setting, and Participants Dietary consumption and follow-up data from 84 688 female nurses enrolled in the Nurses' Health Study, aged 34 to 59 years and free from cardiovascular disease and cancer at baseline in 1980, were compared from validated questionnaires completed in 1980, 1984, 1986, 1990, and 1994.

Main Outcome Measures Incident nonfatal myocardial infarction and CHD deaths.

Results During 16 years of follow-up, there were 1513 incident cases of CHD (484 CHD deaths and 1029 nonfatal myocardial infarctions). Compared with women who rarely ate fish (<1 per month), those with a higher intake of fish had a lower risk of CHD. After adjustment for age, smoking, and other cardiovascular risk factors, the multivariable relative risks (RRs) of CHD were 0.79 (95% confidence interval [CI], 0.64-0.97) for fish consumption 1 to 3 times per month, 0.71 (95% CI, 0.58-0.87) for once per week, 0.69 (95% CI, 0.55-0.88) for 2 to 4 times per week, and 0.66 (95% CI, 0.50-0.89) for 5 or more times per week (P for trend = .001). Similarly, women with a higher intake of omega-3 fatty acids had a lower risk of CHD, with multivariable RRs of 1.0, 0.93, 0.78, 0.68, and 0.67 (P<.001 for trend) across quintiles of intake. For fish intake and omega-3 fatty acids, the inverse association appeared to be stronger for CHD deaths (multivariate RR for fish consumption 5 times per week, 0.55 [95% CI, 0.33-0.90] for CHD deaths vs 0.73 [0.51-1.04]) than for nonfatal myocardial infarction.

Conclusion Among women, higher consumption of fish and omega-3 fatty acids is associated with a lower risk of CHD, particularly CHD deaths.

Low rates of cardiovascular disease in populations with a high intake of fish, such as Alaskan Natives,^1,2 Greenland Eskimos,^3,4 and Japanese people residing in fishing villages,^5,6 have suggested that fish consumption may protect against atherosclerosis. Several^7-9 but not all prospective cohort studies^10,11 have found an inverse association between fish consumption and risk of coronary heart disease (CHD). In addition, 2 secondary-prevention trials^12,13 showed that increasing fish consumption or fish-oil supplementation reduced coronary mortality among patients with preexisting coronary disease. Virtually all previous studies on fish consumption and CHD were conducted in men; these results may not apply to women. We therefore examined the association between fish and long-chain omega-3 fatty acid intake and incidence of CHD among women in the Nurses' Health Study cohort during 16 years of follow-up.

The Nurses' Health Study began in 1976, when 121 700 women who were registered nurses (98% white) aged 30 to 55 years and living in 11 US states completed questionnaires about their lifestyle and medical history. Every 2 years, follow-up questionnaires were sent to update information and identify new major illnesses. A total of 98 462 women returned the 1980 dietary questionnaire. We excluded women who left 10 or more items blank, those with reported total food intakes judged to be implausible, and those who had a history of cancer (except nonmelanoma skin cancer), angina, myocardial infarction (MI), coronary revascularization, stroke, or other cardiovascular diseases before 1980. After these exclusions, 84 688 women remained for these analyses.

The semiquantitative food frequency questionnaire used in 1980 included a survey of 61 foods, including a single question assessing fish intake.¹⁴ A common unit or portion size for each food (eg, 168-224 g [6-8 oz] for fish) was specified, and each woman was asked how often on average during the previous year she had consumed that amount. Nine responses were possible for each food item, ranging from "almost never" to "6 or more times per day." In 1984, 1986, 1990, and 1994, the dietary questionnaire was expanded to include 4 fish and seafood items: (1) dark-meat fish such as mackerel, salmon, sardines, bluefish, or swordfish (84-140 g [3-5 oz]); (2) canned tuna (84-112 g [3-4 oz]); (3) other fish (84-140 g [3-5 oz]); and (4) shrimp, lobster, or scallops as the main dish (98 g [3.5 oz]). The average daily intake of nutrients was calculated by multiplying the frequency of consumption of each item by its nutrient content per serving and totaling the nutrient intake for all food items.

The calculation of long-chain omega-3 fatty acids has been described in detail elsewhere.¹⁵ Briefly, to calculate intake of omega-3 fatty acids (eicosapentaenoic acid and docosahexanoic acid), we assigned grams per serving as follows: 1.51 for dark-meat fish, 0.42 for canned tuna fish, 0.48 for other fish, and 0.32 for shrimp, lobster, or scallops. These omega-3 fatty acid values were derived by weighting the mean values of omega-3 fatty acids for the most common types of fish according to US landing data in 1984 (US Department of Commerce). To make the intake of marine omega-3 fatty acids from the 1980 questionnaire as comparable as possible with the later, more detailed questionnaires, we assigned 1.16 g of long-chain omega-3 fatty acids per portion (168-224 g [6-8 oz]) on the 1980 questionnaire. This number was calculated as a weighted average of omega-3 fatty acid composition from dark-meat fish, canned tuna, and other fish by using the relative consumption of these types of fish on the 1984 dietary questionnaire. Intake of long-chain omega-3 fatty acids was primarily from fish (87% of the total intake) and secondarily from chicken (7%) and liver (2%), which is similar to that in the US food supply data.¹³

The reproducibility and validity of the 1984 dietary questionnaire were assessed in a random sample of 127 men aged 45 to 70 years and living in the Boston area by comparing the data from the questionnaire with the data from 2 one-week dietary records, collected approximately 6 to 8 months apart,¹⁶ and with the fatty acid composition of adipose tissue.¹⁷ Spearman rank correlation coefficients for the fish items between 2 questionnaires administered 1 year apart were 0.63 for dark-meat fish, 0.54 for canned tuna, 0.48 for other fish, and 0.67 for shrimp, lobster, or scallops as a main dish.¹⁶ The mean total fish intake was 3.7 servings weekly according to the questionnaire and 3.6 servings weekly according to 2 one-week dietary records (Spearman rank correlation coefficient, 0.61; P<.001). The energy-adjusted intake of eicosopentaenoic acid from fish also was correlated with the percentage of eicosopentaenoic acid in adipose tissue (Spearman rank correlation, 0.49; P<.001).¹⁷ Information on fish oil supplements was not requested until 1990 in the Nurses' Health Study; at that point, the prevalence of consumption of this supplement was only 1.6%.

The end point for this study was incidence of CHD (including CHD deaths and nonfatal MI) occurring after return of the 1980 questionnaire but before June 1, 1996. We sought to review medical records for all self-reported MIs. Records were reviewed by physicians with no knowledge of the self-reported risk factor status. The diagnosis of MI was confirmed by using World Health Organization criteria: symptoms in addition to either diagnostic electrocardiographic changes or elevated cardiac enzyme levels.¹⁸ Infarctions that required hospital admission and for which confirmatory information was obtained by interview or letter, but for which no medical records were available, were designated as probable (17%).

Deaths were identified from state vital records and the National Death Index or reported by next of kin and the postal system. Follow-up for death was more than 98% complete.¹⁹ Fatal coronary disease was defined as fatal MI if it was confirmed by hospital records or autopsy, if coronary disease was listed as the cause of death on the death certificate and was the underlying and most plausible cause, and if evidence of previous coronary disease was available. We designated as presumed coronary disease (15% fatal cases) those in which coronary disease was the underlying cause on the death certificate but for which no records were available.

For each woman, person-months of follow-up were calculated from the date of return of the 1980 questionnaire to the first end point, death, or June 1, 1996, whichever came first. Women who reported having cardiovascular disease or cancer on previous questionnaires were excluded from subsequent follow-up.

Because of the long follow-up period, dietary variables were updated to better represent long-term dietary patterns, using the information from 1980, 1984, 1986, 1990, and 1994 dietary questionnaires. We calculated intake of fish and omega-3 fatty acids as a cumulative average of intake from all available dietary questionnaires up to the start of each 2-year follow-up interval in which events were reported.²⁰ Because changes in diet after the development of intermediate end points such as angina, hypercholesterolemia, and diabetes may confound information on diet and disease, we stopped updating information on diet at the beginning of the interval during which these intermediate end points developed in an individual subject.²¹ The other nutrient variables (fiber, trans-fat, and the ratio of polyunsaturated to saturated fats) and intake of fruits and vegetables and red meat (beef, pork, or lamb as the main dish or mixed dish) were also calculated as a cumulative average of intake.

We divided women into 5 categories according to frequency of fish consumption (<1 per month, 1-3 times per month, 1 per week, 2-4 times per week, and ≥5 times per week) or quintiles of omega-3 fatty acids (as the percentage of total energy) and calculated incidence rates by dividing the number of events by person-time of follow-up in each category. The relative risk (RR) was computed as the rate in a specific category of fish or omega-3 fatty acid consumption divided by that in the lowest category, with adjustment for age in 5-year categories.

We used Cox proportional hazards modeling (PROC PHREG) for all multivariable analyses (SAS Institute Inc, Cary, NC). The Anderson-Gill data structure was used to handle time-varying covariates efficiently,²² where a new data record was created for every questionnaire cycle at which a participant was at risk, with covariates set to their values when the questionnaire was returned. To control as finely as possible for confounding by age, calendar time, and any possible 2-way interactions between these 2 time scales, we stratified the analysis jointly by age in months at the start of follow-up and calendar year of the current questionnaire cycle. The time scale for the analysis was measured as months since the start of the current questionnaire cycle, which is equivalent to age in months. In the multivariable models, we simultaneously included total energy intake; cigarette smoking; body mass index; menopausal status and postmenopausal hormone use; alcohol use; history of hypertension, high cholesterol, and diabetes; multivitamin use; vitamin E supplement use; family history of MI; physical activity (number of hours spent on moderate vigorous exercise per week); and aspirin use (Table 1). In a second multivariate model, we also adjusted for intake of trans-fat, fiber, and the ratio of polyunsaturated to saturated fats. Because further adjustment for intake of fruits and vegetables (5 categories), red meat (quintiles), or α-linolenic acid did not appreciably alter the results, we did not include them in the final model.

We conducted analyses stratified by aspirin use and the ratio of omega-6 polyunsaturated fat (linoleic acid) to omega-3 fatty acids (the sum of α-linolenic acid, eicosapentaenoic acid, and docosahexaenoic acid) to assess possible effect modification by these variables. We tested the significance of the interaction with a likelihood ratio test by comparing a model with the main effects of omega-3 fatty acids and the stratifying variable and the interaction terms with a reduced model with only the main effects.

In a supplemental analysis, we performed a propensity analysis²³ in which we used logistic regression modeling to describe the likelihood of being a woman who consumes fish frequently as opposed to rarely. Demographic, clinical, and nonfish dietary variables were included in the propensity model. We used the resulting propensity scores to match women from the 2 groups.

During 16 years of follow-up (1 307 157 person-years), we documented 1513 incident cases of CHD (484 CHD deaths and 1029 nonfatal MIs). As described in detail elsewhere,¹⁵ compared with women who seldom ate fish, women with a higher fish consumption were slightly older, had a lower prevalence of current smoking, and had a higher prevalence of being overweight and of hypertension, vigorous activity, regular aspirin use, and mutivitamin use. Fish consumption was positively associated with intake of chicken, fruits and vegetables, and dairy foods and was inversely associated with intake of red meat.

Table 1 presents relative risk of CHD according to fish intake. We observed significant inverse associations between fish intake and incidence of CHD after adjustment for age alone and age plus standard cardiovascular risk factors (P for trend <.001). After further adjustment for other cardiovascular risk factors, the association was still significant (P for trend = .001). Further adjustment for dietary factors did not appreciably alter the results. The inverse association was somewhat stronger for fatal CHD (multivariate RR comparing extreme categories of fish consumption was 0.55 [95% CI, 0.33-0.90]) than for nonfatal MI (corresponding RR was 0.73 [95% CI, 0.51-1.04]).

Intake of dietary omega-3 fatty acids was significantly inversely associated with a lower risk of CHD (Table 2) (P for trend <.001). Further adjustment for dietary factors did not materially change these RRs. Again, this inverse association appeared to be stronger for fatal CHD than for nonfatal MI (Table 2).

The inverse association between intake of fish and omega-3 fatty acids and risk of CHD was stronger for women who did not use aspirin regularly (<1/week) than for regular aspirin users (Table 3), but tests for interaction did not reach statistical significance (P for interaction between fish consumption and aspirin use = .14; P for interaction between omega-3 fatty acids and aspirin use = .27). Intake of omega-3 fatty acids was significantly associated with a lower risk of CHD death among regular aspirin users and nonusers. In contrast, the inverse association between omega-3 fatty acids and nonfatal MI was significant only among women who did not use aspirin regularly (RR comparing extreme quintiles = 0.68; 95% CI, 0.42-1.09; P for trend = .008), not among regular aspirin users (RR comparing extreme quintiles = 0.83; 95% CI, 0.59-1.16; P for trend = .15).

We conducted a stratified analysis according to the ratio of omega-6 (linoleic) to omega-3 (the sum of α-linolenic acid and fish oil) fatty acid intake. In this analysis, we used the mean of the omega-6/omega-3 ratio (7.6) as a cutoff point to define low and high ratio groups. As shown in Table 4, omega-3 fatty acids were associated with a significantly lower risk of CHD in both low- and high-ratio omega-6/omega-3 groups. The inverse association appeared to be somewhat stronger in the high-ratio group than in the low-ratio group, but test for interaction was not statistically significant (P = .38).

During the 16 years of follow-up, we documented 4121 deaths from all causes in the cohort. Both fish and omega-3 fatty acid consumption were associated with a lower risk of all-cause mortality. The multivariate RR of all-cause mortality comparing women consuming fish at least 5 times weekly with those consuming fish less than once a month was 0.68 (95% CI, 0.57-0.82; P for trend <.001). The multivariate RR of all-cause mortality comparing extreme quintiles of omega-3 fatty acid intake was 0.75 (95% CI, 0.67-0.85; P for trend <.001).

In the supplementary analysis of propensity-matched women, those who consumed fish frequently still had a reduced risk of CHD compared with those who ate fish rarely (adjusted RR, 0.62; 95% CI, 0.41-0.92).

In this prospective cohort study, we observed a significant inverse association between fish and omega-3 fatty acid consumption and incidence of major CHD events, CHD deaths in particular, throughout a 16-year follow-up. This inverse association was independent of established cardiovascular risk factors and dietary predictors of CHD such as fiber, trans-fatty acids, and the ratio of polyunsaturated to saturated fats. It was also not explained by differences in intake of red meat or fruits and vegetables.

Several prospective cohort studies examined associations between fish consumption and risk of CHD in men. In the Dutch component of the Seven Countries Study, with 20 years of follow-up, Kromhout et al⁷ found that men who consumed 30 g of fish daily had a 50% lower CHD mortality than men who rarely ate fish. In the Western Electric Study, Daviglus et al⁸ found that men who consumed at least 35 g of fish daily had a 40% lower risk of fatal CHD. In the US Physicians' Health Study, Albert et al⁹ found that weekly fish consumption was associated with a lower risk of sudden cardiac death (RR, 0.48; 95% CI, 0.24-0.96) but was not related to risks of nonfatal MI or other cardiovascular end points.¹⁰ In the Health Professionals Follow-up Study, Ascherio et al¹¹ found no overall association between dietary intake of omega-3 fatty acids or fish intake and the risk of coronary disease, but there was a nonsignificant trend for a lower risk of fatal CHD with increasing fish consumption. Our study is broadly consistent with previous studies, suggesting that fish consumption is probably more protective against fatal CHD than nonfatal MI. Notably, previous studies were conducted in men, none used repeated measures of fish intake in the analysis, and few adjusted for potential confounding dietary variables.

Two secondary prevention trials, the Diet and Reinfarction Trial (DART)¹² and the GISSI-Prevenzione trial,¹³ showed that fish consumption or fish oil supplementation also reduces coronary mortality among patients after MI. In the DART, which included 2033 men allocated to 3 dietary interventions, subjects who received advice to eat more fish had a significantly lower (29%) total mortality during 2 years of follow-up. There was also a nonsignificant trend toward a reduction in recurrent ischemic heart disease events with increased fatty fish consumption. In the more recent GISSI-Prevenzione trial, which included 11 324 MI patients (primarily men), daily supplementation (1 g/d) of omega-3 fatty acids for 2 years reduced occurrence of the main cardiovascular end points (cardiovascular death, nonfatal MI, and stroke) by 20%, cardiovascular death (including coronary or cardiac deaths and sudden deaths) by 30%, and all fatal events by 20%. In a secondary prevention trial, post-MI patients receiving an experimental diet that included greater intake of α-linolenic acid (a shorter-chain omega-3 fatty acid) experienced a significant reduction in sudden cardiac death and all-cause mortality.^24,25 Our findings that fish and omega-3 fatty acid consumption was associated with a lower risk of overall mortality were consistent with those in other trials.

In our cohort, omega-3 fatty acid intake and fish consumption were associated with a significantly lower risk of CHD. This finding is consistent with the hypothesis that omega-3 fatty acids are the active agent primarily responsible for the apparent protective effect of fish. Omega-3 fatty acids may reduce CHD incidence and mortality through multiple mechanisms, including reduction of serum triglycerides,²⁶ platelet aggregability,²⁷ and antiarrhythmic effects.²⁸ Animal studies have established that fish oil intake effectively reduces the incidence and duration of cardiac arrhythmia.²⁹ In a case-control study, Siscovick et al³⁰ found that dietary intake and cell membrane levels of long-chain omega-3 fatty acids were associated with a significantly lower risk of primary sudden cardiac arrest. Multiple mechanisms have been proposed to explain the antiarrhythmic effect of fish oil, including modification of the eicosanoid system (eg, reducing the production of thromboxane A₂), alteration of the fatty acid composition of membrane phospholipids, effects on various enzymes and receptors,³¹ inhibition of the voltage-gated sodium channels, and changes in heart rate variability.³²

There is growing evidence to support the hypothesis that fish oil improves endothelial dysfunction, which is considered an early marker of atherosclerosis.^33,34 In vitro studies have consistently shown that omega-3 fatty acids decrease expression of adhesion molecules on the endothelium and also decrease leukocyte-endothelium interactions.³³ Additionally, clinical experimental studies have shown that omega-3 fatty acid supplementation improves endothelial-dependent vasomotor function.^34,35 The beneficial effects of omega-3 fatty acids on endothelial function may in part explain the inverse association we observed for nonfatal MI.

Aspirin decreases the risk of CHD in part by blocking the cyclooxygenase enzyme that converts arachidonic acid to thromboxane, inhibiting platelet aggregation.³⁶ Long-chain omega-3 fatty acids also interact with cyclooxygenase enzymes; however, these fatty acids act as a substrate for the enzymes, leading to the production of an inactive thromboxane moiety in platelets and an active prostaglandin molecule in endothelial cells.³⁷ The result is reduced platelet aggregation and increased dilation of blood vessels. Since aspirin is a more potent inhibitor of cyclooxygenase enzymes than fish oil, it is possible that aspirin use could mask the effect of small amounts of fish intake. Consistent with our a priori hypothesis, the inverse association between omega-3 fatty acids and MI was stronger among women who did not use aspirin regularly than among regular users, which is consistent with our previous analyses of fish consumption and ischemic stroke.¹⁵ However, even among regular aspirin users, a higher omega-3 fatty acid intake was associated with a significant reduction in CHD death, suggesting that mechanisms other than those related to prostaglandin metabolism were involved.

Because omega-6 and omega-3 fatty acids compete for delta-6 desaturase enzyme in the desaturation and chain elongation pathway and a higher intake of omega-6 fatty acid may lead to an increase in the production of thromboxane A₂, a proaggregatory vasoconstrictor,^31,38 a higher ratio of omega-6 to omega-3 fatty acid intake may attenuate the benefit of omega-3 fatty acid. In our study, however, the inverse association between omega-3 fatty acid and CHD persisted in women with a low or high ratio of omega-6/omega-3 fatty acid intake. Clearly, reducing thrombotic tendency is only one of several biological mechanisms through which long-chain omega-3 fatty acids may lower the risk of CHD. Because a higher intake of linoleic acid²⁰ and α-linolenic acid³⁹ was associated with a lower risk of CHD in our cohort, the ratio of omega-6 to omega-3 fatty acids was not appreciably associated with CHD risk.

Our large sample size and long follow-up provided adequate power to look at CHD deaths and nonfatal MIs separately and conduct planned subgroup analyses. A unique advantage of this study is that fish consumption was assessed multiple times, and our analyses using cumulative averages not only took into account changes in eating behaviors, but also reduced measurement errors caused by intrasubject variation.²¹ Finally, we were able to adjust for important nondietary and dietary covariates, which were also updated over time.

These observational data cannot prove that fish consumption causes a reduction in CHD risk. In our cohort, women who consumed more fish had a somewhat healthier diet and lifestyle. However, careful adjustment for potential dietary and lifestyle confounding variables did not appreciably alter the results, suggesting an independent effect of fish and omega-3 fatty acids on CHD risk. Still, the possibility of unmeasured or incompletely controlled confounding cannot be excluded. Nevertheless, the biological plausibility of a causal relationship between fish consumption and reduction in CHD risk through antiarrhythmic and antithrombotic effects of fish oil, as well as the consistency of the present findings with those from other prospective cohort studies and secondary prevention trials, supports the likelihood of a causal association.

In conclusion, this prospective study provides strong evidence for an inverse association between fish and omega-3 fatty acid consumption and risk of CHD in women, particularly CHD death. These findings lend further support to current dietary guidelines recommending fish consumption twice weekly for the prevention of CHD.⁴⁰