Fish Intake, Marine Omega-3 Fatty Acids, and Mortality in a Cohort of Postmenopausal Women

Transcription

1 American Journal of Epidemiology Copyright 2004 by the Johns Hopkins Bloomberg School of Public Health All rights reserved Vol. 160, No. 10 Printed in U.S.A. DOI: /aje/kwh307 Fish Intake, Marine Omega-3 Fatty Acids, and Mortality in a Cohort of Postmenopausal Women Aaron R. Folsom and Zewditu Demissie From the Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, MN. Received for publication November 10, 2003; accepted for publication June 2, Intake of fish or omega-3 fatty acids may decrease risk of total and coronary heart disease death, but evidence from low-risk populations is less convincing. The authors assessed intake by using a food frequency questionnaire at baseline in a cohort of Iowa women aged years. Among women initially free of heart disease and cancer (4,653 deaths over 442,965 person-years), there was an inverse age- and energy-adjusted association between total mortality and fish intake, with a relative risk of 0.82 (95% confidence interval: 0.74, 0.91) for the highest versus lowest quintile. Age- and energy-adjusted associations also were inverse (p for trend < 0.05), although not entirely monotonic, for cardiovascular, coronary heart disease, and cancer mortality. Adjustment for multiple other risk factors attenuated all associations to statistically nonsignificant levels. Estimated marine omega-3 fatty acid intake also was not associated with total or cause-specific mortality. In comparison, plant-derived α-linolenic acid was inversely associated with mortality after multivariable adjustment. Intake of neither fish nor marine omega-3 fatty acids was associated with breast cancer incidence. These findings do not argue against recommending fish as part of a healthy diet, as other evidence suggests benefit. Nevertheless, the authors of this study could not verify that fish and marine omega-3 fatty acid intake had independent health benefits in these postmenopausal women. breast neoplasms; coronary disease; fatty acids, omega-3; fishes; mortality; prospective studies Abbreviations: ICD-9, International Classification of Diseases, Ninth Revision; ICD-10, International Classification of Diseases, Tenth Revision. Considerable evidence from both clinical trials and observational studies indicates that greater intake of fish, or of omega-3 fatty acids from fish or plant sources, may decrease risk of coronary heart disease incidence and mortality (1 7), and sudden death in particular (8 10). There is mixed evidence that fish or omega-3 fatty acids may prevent stroke (1). In observational studies, the inverse association of fish or marine omega-3 intake with total coronary heart disease mortality appears to be stronger in populations at high risk of coronary heart disease and may be nonexistent in low-risk populations (2, 4, 11). Fewer studies exist for women than men. In a cohort of women with generally low coronary heart disease mortality, we assessed whether fish or marine omega-3 fatty acid intake is inversely associated with major causes of death. MATERIALS AND METHODS The Iowa Women s Health Study cohort comprises 41,836 women aged years recruited via a baseline questionnaire mailed in A previous report explained how selfreported baseline risk factors were assessed and defined (12). Baseline histories of physician-diagnosed cancer, heart attack, angina, or other heart disease were also obtained. We assessed baseline dietary intake by using a 127-item food frequency questionnaire. Four fish and seafood questions asked about frequency of intake of 1) dark-meat fish such as mackerel, salmon, sardines, bluefish, or swordfish ( g (3 5 ounces)); 2) canned tuna ( g (3 4 ounces)); 3) other fish ( g (3 5 ounces)); and 4) shrimp, lobster, or scallops as a main dish (98 g (3.5 ounces)). Frequency categories ranged from never or less than once per month to 6 or more per day. We derived Correspondence to Dr. Aaron R. Folsom, Division of Epidemiology, School of Public Health, University of Minnesota, Suite 300, 1300 South Second Street, Minneapolis, MN ( folsom@epi.umn.edu). 1005

2 1006 Folsom and Demissie total fish and seafood servings from a weighted average of the answers to the four seafood questions plus any information about fish or seafood entered in the other foods eaten at least once per week section of the questionnaire. We calculated the average daily intake of omega-3 fatty acids from fish and other nutrients by multiplying the frequency of consumption of each item by its nutrient content per serving and totaling the nutrient intake for all food items. Omega-3 fatty acid content was estimated as described by Hu et al. (13). We did not ask about fish oil supplements, but the Nurses Health Study investigators reported rare use in their cohort (1.6 percent in 1990) (13). In other populations, fish intake according to this food frequency questionnaire has correlated moderately highly (r = ) with marine omega-3 fatty acid content of tissues (14, 15). We identified cancer incidence and most deaths by annual linkage of cohort identifiers to Iowa State wide cancer incidence and death records and by questionnaires mailed to the cohort in 1987, 1989, 1992, and To find additional deaths, we sent identifiers of nonrespondents to follow-up surveys to the National Death Index. Cause of death was that assigned as the underlying cause by state health departments, as follows: cardiovascular (International Classification of Diseases, Ninth Revision (ICD-9) codes ; International Classification of Diseases, Tenth Revision (ICD-10) codes I00 I99), coronary heart disease (ICD-9 codes , 429.2; ICD-10 codes I20 I25, I51.6), stroke (ICD-9 codes ; ICD-10 codes I60 I69, G45), and cancer (ICD-9 codes ; ICD-10 codes C00 D48). We did not have data on suddenness of death. We used quintiles to analyze our single 1986 assessment of total fish and seafood intake and of marine omega-3 intake. We used analysis of covariance to examine the relation of fish quintiles to other risk factors. We used Poisson regression or proportional hazards models to compute relative risks and 95 percent confidence intervals of death, adjusted for 1) age and energy and 2) covariates previously reported to be associated with total and cardiovascular mortality in this cohort (12). Follow-up time for mortality was from baseline until death or else December 31, For breast cancer incidence, women were followed up until breast cancer occurrence, emigration from Iowa, death, or else December 31, A test for trend in relative risks was performed by using an ordinal variable in the model to represent each quintile. RESULTS Fish intake at baseline was similar to that in most contemporary US populations studied by using a comparable food frequency questionnaire (7, 10). The respective proportions of women who reported eating a serving of fish less than once per month, 1 3 times per month, or 1 or more times per week were 67 percent, 24 percent, and 9 percent for darkmeat fish; 30 percent, 40 percent,and 30 percent for tuna; 35 percent, 36 percent, and 29 percent for other fish; and 79 percent, 18 percent, and 3 percent for shrimp, lobster, or scallops. The mean respective intakes of eicosapentaenoic acid, docosohexaenoic acid, and total marine omega-3 fatty acids were 53 mg, 135 mg, and 188 mg per day. Mean intake of α-linolenic fatty acid, a nonmarine omega-3 fatty acid, was 1.09 g per day. Women initially free of heart disease As table 1 shows, greater fish intake was associated with younger age; with greater education, physical activity, alcohol consumption, estrogen use, vitamin use, body mass index, and hypertension; but with slightly less smoking. Greater fish intake also was associated with greater intake of energy and most other foods or nutrients examined. There were 4,653 deaths during 442,965 person-years of follow-up in women initially free of heart disease and cancer. When adjusted for age and energy intake, the association between total mortality and baseline fish intake was inverse (p for trend = 0.003), with a relative risk of 0.82 (95 percent confidence interval: 0.74, 0.91) for the highest versus lowest quintile (table 2). Age- and energy-adjusted associations also were inverse (p for trend < 0.05), although not entirely monotonic, for cardiovascular, coronary heart disease, and cancer mortality. In contrast, fish intake was not associated with stroke mortality; noncardiovascular, noncancer deaths; or breast cancer incidence. Adjustment for multiple other risk factors attenuated all associations to statistically nonsignificant levels (table 2). For example, the multivariately adjusted relative risk of total mortality was 0.93 (95 percent confidence interval: 0.83, 1.05) for quintile 5 versus 1 of fish intake (p for trend = 0.15). None of the four specific categories of fish listed on the food frequency questionnaire showed an independent association with mortality either. Likewise, in analyses restricted to diabetic women, there was no evidence of an association of fish intake with mortality from all causes (age- and energy-adjusted relative risks across quintiles = 1.0, 0.97, 0.90, 1.05, 0.92; p for trend = 0.78), cardiovascular disease (relative risks = 1.0, 0.86, 0.84, 1.03, 0.91; p for trend = 0.82), or coronary heart disease (relative risks = 1.0, 0.98, 0.93, 1.28, 1.16; p for trend = 0.68). Among women initially free of heart disease and cancer, estimated marine omega-3 fatty acid intake was not associated with total mortality or breast cancer incidence in reduced or more completely adjusted models (table 3). This finding was true for cause-specific deaths as well, and for diabetic and nondiabetic women (results not shown). Total mortality also was not associated separately with eicosapentaenoic acid tertiles (age- and energy-adjusted relative risks = 1.0, 0.99, 0.98; p for trend = 0.91) or docosahexaenoic acid tertiles (age- and energy-adjusted relative risks = 1.0, 0.98, 0.95; p for trend = 0.41). In a supplemental analysis, however, plant-derived α-linolenic acid was modestly inversely associated with total mortality (relative risks across tertiles = 1.0, 0.95, 0.85; p for trend = 0.01, adjusted for all covariates). Women with a history of heart disease As a secondary analysis (not shown in the tables), we also examined the association of total mortality with fish intake for women who at baseline were free of cancer but reported a history of myocardial infarction, angina, or other heart

3 Fish Intake and Mortality in Older Women 1007 TABLE 1. Distribution of baseline risk factors in relation to baseline fish intake among participants initially free of cancer and cardiovascular disease, Iowa Women s Health Study, 1986 Prevalence (%) * Reflects the serum-cholesterol-raising capacity of the diet. Frequency of fish intake per week (servings, approximate quintiles) < to < to 1.5 >1.5 to < p for trend Age >62 years < Education high school < Low level of physical activity < Alcohol nonconsumer < Current smoker First livebirth at age 30 years Current estrogen user < Vitamin user < Body mass index >30 kg/m < Waist/hip ratio > Diabetes Hypertension < Mean Energy intake (kcal/day) 1,607 1,673 1,797 1,888 1,973 < Whole-grain intake (servings/week) < Fruit and vegetables intake (servings/week) < Red meat intake (servings/week) < Keys score* < Cholesterol (mg/day) < Saturated fat intake (g/day) < α-linolenic intake (g/day) < disease (1,069 deaths, 42,095 person-years). We found a modest, inverse association between fish intake and total mortality in these women. The age- and energy-adjusted relative risks of total mortality across quintiles of fish intake were 1.00, 1.09 (95 percent confidence interval: 0.88, 1.36), 0.95 (95 percent confidence interval: 0.78, 1.15), 0.83 (95 percent confidence interval: 0.64, 1.07), and 0.88 (95 percent confidence interval: 0.71, 1.10); p for trend = This association was eliminated with multivariate adjustment (p for trend = 0.88). Estimated marine omega-3 fatty acid intake and specific groups of fish or seafood also were unrelated to total mortality (p for trend = 0.85) in women with a history of heart disease. DISCUSSION In this prospective study of older Iowa women, greater fish intake was associated with modestly reduced mortality from all causes, coronary heart disease, and cancer. However, the association was not independent of other self-reported risk factors. Intake of marine omega-3 fatty acids, the presumed beneficial nutrient in fish (specifically, eicosapentaenoic and docosahexaenoic acids) (1, 3), showed no association with total or cause-specific mortality. Unlike a previous study (13), we found no associations between fish or omega-3 fatty acids and mortality among diabetic women either. There also was no association of fish intake with breast cancer incidence. An inverse association between fish or omega-3 fatty acids and coronary heart disease or stroke has been shown by many, but not all, previous cohort studies and in trials using diet or fish oil capsules (1 3). The association is biologically plausible; fish or omega-3 fatty acids in adequate amounts can be antiinflammatory, hypotriglyceridemic, antithrombotic, ventricular antiarrhythmic, endothelium relaxant, and possibly antiatherogenic (1). In addition, greater fish intake often replaces other atherogenic components in the diet. In contrast to our null findings for fish, α-linolenic acid from plants did show an inverse association with total mortality in a supplemental analysis, consistent with previous reports (1, 3). There is little prior evidence for greater fish intake preventing cancer overall (16), although recent data suggest a potential inverse association for breast cancer (17). We could not confirm an association with breast cancer incidence. Other studies should examine the relation of fish intake with breast cancer.

4 1008 Folsom and Demissie TABLE 2. Relative risks of total and cause-specific deaths and incident breast cancer in relation to baseline fish intake among participants initially free of cancer and cardiovascular disease, Iowa Women s Health Study, Frequency of fish intake per week (servings, approximate quintiles) < to < to 1.5 >1.5 to < Mortality person-years 50,038 77, ,852 48,325 92,341 Total mortality (no. of events) p for trend RR1* % CI Reference 0.8, , , , 0.91 RR % CI Reference 0.88, , , , 1.05 Cardiovascular disease mortality (no. of events) RR1* % CI Reference 0.77, , , , 1.03 RR % CI Reference 0.85, , , , 1.15 Coronary heart disease mortality (no. of events) RR1* % CI Reference 0.78, , , , 1.20 RR % CI Reference 0.87, , , , 1.34 Stroke mortality (no. of events) RR1* % CI Reference 0.82, , , , 1.53 RR % CI Reference 0.86, , , , 1.67 Cancer mortality (no. of events) RR1* % CI Reference 0.74, , , , 0.95 RR % CI Reference 0.81, , , , 1.11 Other mortality (no. of events) RR1* % CI Reference 0.75, , , , 0.96 RR % CI Reference 0.77, , , , 1.17 Breast cancer incidence (no. of events) Person-years 47,369 72, ,196 45,162 86,143 RR1* % CI Reference 0.87, , , , 1.20 RR % CI Reference 0.80, , , , 1.12 * Relative risk (RR) adjusted for age (continuous) and energy intake (quintiles). CI, confidence interval. Relative risk adjusted for age, energy intake, educational level (<high school, high school, or >high school), physical activity level (low, medium, or high), alcohol consumption (0, <4, or 4 g/day), smoking status (current, former, or never), pack-years of cigarette smoking (continuous), age at first livebirth (nullipara, <30 years, or 30 years), estrogen use (current, former, or never), vitamin use (yes, no, or unknown), body mass index (quintiles), waist/hip ratio (quintiles), diabetes (yes or no), hypertension (yes, no, or unknown), intake of whole grains, fruit and vegetables, red meat, cholesterol, and saturated fat (all in quintiles). Few previous studies of fish intake and mortality have focused on women (1). The most notable positive study was the Nurses Health Study, which showed an approximately one-third lower coronary heart disease risk for fish eating at

5 Fish Intake and Mortality in Older Women 1009 TABLE 3. Relative risks of total mortality and breast cancer incidence in relation to baseline quintiles of estimated omega-3 fatty acids from fish among participants initially free of cancer or cardiovascular disease, Iowa Women s Health Study, Quintile of omega-3 fatty acid intake (g/day) Mean intake Total mortality Person-years 86,882 82,248 96,107 89,342 88,387 No. of events , p for trend RR1* % CI Reference 0.85, , , , 1.00 RR % CI Reference 0.89, , , , 1.06 Breast cancer incidence Person-years 81,929 77,150 90,043 83,798 82,759 No. of events RR1* % CI Reference 0.90, , , , 1.17 RR % CI Reference 0.84, , , , 1.08 * Relative risk (RR) adjusted for age (continuous) and energy intake (quintiles). CI, confidence interval. Relative risk adjusted for age, energy intake, educational level (<high school, high school, or >high school), physical activity level (low, medium, or high), alcohol consumption (0, <4, or 4 g/day), smoking status (current, former, or never), pack-years of cigarette smoking (continuous), age at first livebirth (nullipara, <30 years, or 30 years), estrogen use (current, former, or never), vitamin use (yes, no, or unknown), body mass index (quintiles), waist/hip ratio (quintiles), diabetes (yes or no), hypertension (yes, no, or unknown), intake of whole grains, fruit and vegetables, red meat, cholesterol, and saturated fat (all in quintiles). least once per week versus less than once per month (7). The association was stronger for coronary heart disease deaths than for nonfatal myocardial infarction. In contrast, all-cause mortality and coronary heart disease incidence were unrelated to fish intake in a recent cohort of Danish women (11). Likewise, among women in the First National Health and Nutrition Examination Study Epidemiologic Followup Study, there was little relation of fish intake with coronary heart disease or cardiovascular mortality but a borderline inverse association with total mortality (18), as observed here. Interestingly, with respect to sudden death, omega-3 fatty acid supplementation has been shown to improve heart rate variability indices in men but not in women (19). The absence of a significant, independent association between fish intake and total or coronary heart disease mortality in this sample could have several explanations. First, low statistical power to identify meaningful relative risks was not an issue because the population size and number of deaths was large. Second, the fish intake pattern for Iowa women may not have been optimal to test the hypothesis. The median fish intake of one serving per week was similar to many studies showing potential benefit (2, 7, 10). Nevertheless, the range of fish intake may have been too narrow to precisely assess the relative risk or too high in the reference group to estimate the most relevant relative risk (i.e., some fish intake vs. none (1 4)). Furthermore, the most commonly eaten fish appeared to be tuna, probably canned, which is lower in omega-3 fatty acid than is dark, fatty marine fish (1, 3). Third, as with the majority of similar studies, we had a single, self-reported measure of usual fish intake. Errors of recall of diet or changes in dietary intake over time may have obscured any association between fish intake and mortality. The Nurses Health Study updated dietary intake throughout follow-up and observed an inverse association between fish consumption and validated coronary heart disease mortality (7). Use of fish oil capsules was not assessed but is unlikely to have been common enough in this population to have impacted our results (13). Fourth, we studied mortality, not incidence. It is possible that associations between fish intake and mortality do not reflect those for first events. However, associations of fish intake with coronary heart disease in previous studies have been as strong for mortality as for incidence (1). We also had no data on the suddenness of death, which may particularly be reduced by fish or omega-3 fatty acid intake (1, 8 10). Fifth, because we did not validate causes of death, the underlying cause could have been misclassified somewhat. However, an underlying cause of death for coronary heart disease has been shown to be reasonably valid (20). Finally, it could be that there truly is no independent association of fish intake with coronary heart disease mortality in this generally low-risk population (coronary death rate = 2 per 1,000

6 1010 Folsom and Demissie person-years). In contrast to high-risk populations (2), some other studies in low-risk populations (coronary death rates 2 per 1,000 person-years) also have found no association between fish intake and total or coronary heart disease mortality (4, 11). However, even among women who selfreported a prior history of myocardial infarction, angina, or other heart disease in this study, who had a coronary heart disease mortality rate of 8 per 1,000 person-years, we found no association between fish intake and mortality endpoints. A recent secondary prevention trial of omega-3 fatty acids also showed no benefit in patients with angina (21), in contrast with earlier secondary prevention trials showing benefit in patients with a prior myocardial infarction (22, 23). In conclusion, in this sample of postmenopausal women, greater fish intake was weakly, but not independently associated with a reduced rate of death. There was also no independent association of fish intake with coronary heart disease or stroke mortality. These findings do not argue against recommending fish as part of a healthy diet, as other evidence suggests benefit. Nevertheless, we could not verify that fish and marine omega-3 fatty acid intake had independent health benefits in these postmenopausal women. ACKNOWLEDGMENTS This work was supported by National Cancer Institute grant CA REFERENCES 1. Kris-Etherton PM, Harris WS, Appel LJ. Fish consumption, fish oil, omega-3 fatty acids, and cardiovascular disease. Circulation 2002;106: Marckmann P, Grønbaek M. Fish consumption and coronary heart disease mortality. A systematic review of prospective cohort studies. Eur J Clin Nutr 1999;53: Harper CR, Jacobson TA. The fats of life: the role of omega-3 fatty acids in the prevention of coronary heart disease. Arch Intern Med 2001;161: Marckmann P. Fishing for heart protection. Am J Clin Nutr 2003;78: Lemaitre RN, King IB, Mozaffarian D, et al. N-3 polyunsaturated fatty acids, fatal ischemic heart disease, and nonfatal myocardial infarction in older adults: the Cardiovascular Health Study. Am J Clin Nutr 2003;77: Kromhout D, Bosschieter EB, de Lezenne Coulander C. The inverse relation between fish consumption and 20-year mortality from coronary heart disease. N Engl J Med 1985;312: Hu FB, Bronner L, Willett W, et al. Fish and omega-3 fatty acid intake and risk of coronary heart disease in women. JAMA 2002;287: Leaf A, Kang JX, Xiao YF, et al. Clinical prevention of sudden cardiac death by n-3 polyunsaturated fatty acids and mechanism of prevention of arrhythmias by n-3 fish oils. Circulation 2003;107: Albert CM, Campos H, Stampfer MJ, et al. Blood levels of long-chain n-3 fatty acids and the risk of sudden death. N Engl J Med 2002;346: Albert CM, Hennekens CH, O Donnell CJ, et al. Fish consumption and risk of sudden death cardiac death. JAMA 1998;279: Osler M, Andreasen AH, Hoidrup S. No inverse association between fish consumption and risk of death from all-causes, and incidence of coronary heart disease in middle-aged, Danish adults. J Clin Epidemiol 2003;56: Folsom AR, Kushi LH, Anderson KE, et al. Associations of general and abdominal obesity with multiple health outcomes in older women: The Iowa Women s Health Study. Arch Intern Med 2000;160: Hu FB, Cho E, Rexrode KM, et al. Fish and long-chain omega- 3 fatty acid intake and risk of coronary heart disease and total mortality in diabetic women. Circulation 2003;107: Ma J, Folsom AR, Shahar E, et al. Plasma fatty acid composition as an indicator of habitual dietary fat intake in middle-aged adults. Am J Clin Nutr 1995;62: Hunter DJ, Rimm EB, Sacks FM, et al. Comparison of measures of fatty acid intake by subcutaneous fat aspirate, food frequency questionnaire, and diet records in a free-living population of US men. Am J Epidemiol 1992;135: World Cancer Research Fund and the American Institute for Cancer Research. Food, nutrition and the prevention of cancer: a global perspective. Menasha, WI: BANTA Book Group, 1997: Gago-Dominguez M, Yuan JM, Sun CL, et al. Opposing effects of dietary n-3 and n-6 fatty acids on mammary carcinogenesis: The Singapore Chinese Health Study. Br J Cancer 2003;89: Gillum RF, Mussolino M, Madans JH. The relation between fish consumption, death from all causes, and incidence of coronary heart disease: the NHANES I Epidemiologic Follow-up Study. J Clin Epidemiol 2000;53: Christensen JH, Christensen MS, Dyerberg J, et al. Heart rate variability and fatty acid content of blood cell membranes: a dose-response study with n-3 fatty acids. Am J Clin Nutr 1999; 70: White AD, Folsom AR, Chambless LE, et al. Community surveillance of coronary heart disease in the Atherosclerosis Risk in Communities (ARIC) Study: methods and initial two years experience. J Clin Epidemiol 1996;49: Burr ML, Ashfield-Watt PAL, Dunstan DFJ, et al. Lack of benefit of dietary advice to men with angina: results of a controlled trial. Eur J Clin Nutr 2003;57: Burr ML, Fehily AM, Gilbert JF, et al. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: Diet and Reinfarction Trial (DART). Lancet 1989;2: Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI- Prevenzione trial. Gruppo Italiano per lo Studio della Sopravvivenza nell Infarto miocardio. Lancet 1999;354: