[ American Journal of Epidemiology Vol. 161, No. 7Copyright © 2005 by the Johns Hopkins Bloomberg School of Public Health Printed in USA.All rights reserved DOI: 10.1093/aje/kwiOB5Dietary Fat Intake and Risk of Coronary Heart Disease in Women: 20 Yearsof Follow-up of the Nurses' Health StudyKyungwon Oh', Frank B. Hu'2'3, JoAnn E. Manson2'3'4, Meir J. Stampfer'2'3, and Walter C. Willett'2'31 Department of Nutrition, Harvard School of Public Health, Boston, MA.2 Department of Epidemiology, Harvard School of Public Health, Boston, MA.3 The Channing Laboratory, Department of Medicine, Brigham and Women's Hospital and HarvardMedical School, Boston, MA.4 Division of Preventive Medicine, Department of Medicine, Brigham and Women's Hospital and HarvardMedical School, Boston, MA.Received for publication July 16, 2004; accepted for publication November 10, 2004.The authors examined the associations of dietary fat and specific types of fat with risk of coronary heart disease(CHD) among 78,778 US women initially free of cardiovascular disease and diabetes in 1980. They documented1,766 incident CHD cases (including 1,241 nonfatal myocardial infarctions and 525 CHD deaths) during 20 yearsof follow-up. Polyunsaturated fat intake was inversely associated with CHD risk (multivariate relative risk (RR) forthe highest vs. the lowest quintile = 0.75, 95% confidence interval (Cl): 0.60, 0.92; PArend = 0.004), whereas trans-fat intake was associated with an elevated risk of CHD (RR = 1.33, 95% Cl: 1.07, 1.66; PArend = 0.01). Theassociations between intakes of polyunsaturated fat and trans-fat with CHD risk were most evident among womenyounger than age 65 years (for polyunsaturated fat, RR = 0.66, 95% Cl: 0.50, 0.85; PArend = 0.002 and for trans-fat, RR = 1.50, 95% Cl: 1.13, 2.00; A,end = 0.01). The inverse association between polyunsaturated fat intakeand CHD risk was strongest among women whose body mass index was >25 kg/M2.Findings continue to supportan inverse relation between polyunsaturated fat intake and CHD risk, particularly among younger or overweightwomen. In addition, trans-fat intake was associated with increased risk of CHD, particularly for younger women.coronary disease; fatty acids; risk factorsAbbreviations: BMI, body mass index; CHD, coronary heart disease; Cl, confidence interval; RR, relative risk.The results of prospective epidemiologic studies have ingeneral suggested relations between specific types of fatand risk of coronary heart disease (CHD), but no relationwith total fat (1-5). These findings are consistent with theeffects of dietary fats on blood lipids in controlled feedingstudies (6) and the limited randomized trials examiningCHD endpoints (7). In several studies, risk factors (inparticular, blood lipids) for CHD incidence have beenattenuated with advancing age (8-11). Body mass index(BMI) is strongly associated with risk of CHD and is amajor determinant of insulin resistance (12, 13). Themetabolic responses to dietary fat are strongly modified byunderlying insulin resistance (14), but whether the rela-tions between fat intake and risk of CHD are modified byBMI and other risk factors has not been examined indetail.We previously reported the relations of dietary total fatand specific types of fat with risk of CHD over a 14-yearperiod (1). In the present study, we extended this follow-upto 20 years. We hypothesized that the relations betweenspecific types of fat and risk of CHD are stronger amongyounger women. Dietary measurements were made re-peatedly to reduce the impact of errors in dietary assessmentand to account for changes in diet over time.Am J Epidemiol 2005;161:672-679Correspondence to Dr. Walter C. Willett, Department of Nutrition, Harvard School of Public Health, 665 Huntington Avenue, Boston,MA 02115 (e-mail: [email protected]).672Dietary Fat Intake and Coronary Heart Disease in Women 673TABLE 1. Characteristics and risk factors for coronary heart disease according to intake of specific types of fat* In 1990, Nurses'Health Study, United StatesSaturated fat Monounsaturated fat Polyunsaturated fat Trans-fat1 5 1 5 1 5 1 5Age (years) 57 55 57 55 57 55 57 55Body mass index (kg/m2) 24 24 24 24 24 25 24 24Current smoking (%) 14 23 15 22 20 18 16 19Physical activity (hours/week)Y 3.6 2.9 3.7 2.9 3.3 3.0 3.8 2.7History of hypertension (%) 15 16 15 16 16 15 15 16Parental history of MI* (%) 20 19 20 19 20 20 20 19Current hormone use (%) 31 25 29 26 26 30 30 25Aspirin use (%) 51 45 51 46 46 54 50 49Multivitamin use (%) 34 23 35 23 29 28 36 23Total energy (kcal/day) 1,698 1,721 1,684 1,715 1,672 1,764 1,676 1,758Cholesterol (mg/day) 200 244 203 240 219 223 216 222Protein (g/day) 75 76 76 76 76 74 78 73Dietary fiber (g/day) 21 18 21 18 20 19 21 18Alcohol (g/day) 6.5 4.3 6.4 4.1 6.8 4.3 7.1 3.5* 1, lowest quintile; 5, highest quintile.t Hours spent engaging in moderate to vigorous exercise/week.* Ml, myocardial infarction.MATERIALS AND METHODSPopulationThe Nurses' Health Study was initiated in 1976 when121,700 female registered nurses aged 30-55 years com-pleted a mailed questionnaire about their lifestyle factorsand medical history, including previous cardiovasculardisease, cancer, diabetes, hypertension, and high cholesterollevels. Follow-up questionnaires have been sent to thesewomen every 2 years to update information and identifynewly diagnosed major illnesses. A food frequency ques-tionnaire was first administered in 1980. In this analysis, weincluded participants who returned the 1980 questionnaireand excluded those who left 10 or more food items blank orwhose total energy intake was implausible (n = 5,579), andthose who had a history of cardiovascular disease (angina,myocardial infarction, stroke, other cardiovascular disease;n = 1,645), cancer (n = 3,610), diabetes (n = 1,410), orhypercholesterolemia (n = 4,269) before June 1, 1980;women may have changed their diet because of the presenceof these conditions. After these exclusions, data on 78,778women remained in the analysis.Ascertainment of dietA detailed description of our food frequency question-naire and documentation of its reproducibility and validitywere published earlier (1). In 1980, we collected informa-tion on usual diet by using a food frequency questionnaireincluding 61 foods. For each food, a commonly used unit orportion size was specified, and each woman was asked howoften, on average, she had consumed that quantity during theprevious year. Nine responses were