Alcohol Intake and Risk of Coronary Heart Disease in Younger, Middle-Aged,

and Older Adults

Ulla A. Hvidtfeldt, Janne S. Tolstrup, Marianne U. Jakobsen, Berit L. Heitmann,
Morten Grnbk, Eilis O'Reilly, Katarina Blter, Uri Goldbourt, Gran Hallmans,
Paul Knekt, Simin Liu, Mark Pereira, Pirjo Pietinen, Donna Spiegelman, June
Stevens, Jarmo Virtamo, Walter C. Willett, Eric B. Rimm and Alberto Ascherio
Circulation 2010;121;1589-1597; originally published online Mar 29, 2010;
DOI: 10.1161/CIRCULATIONAHA.109.887513
Circulation is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX
72514
Copyright 2010 American Heart Association. All rights reserved. Print ISSN: 0009-7322. Online
ISSN: 1524-4539

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://circ.ahajournals.org/cgi/content/full/121/14/1589

Subscriptions: Information about subscribing to Circulation is online at

http://circ.ahajournals.org/subscriptions/

Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax:
410-528-8550. E-mail:
journalpermissions@lww.com

Reprints: Information about reprints can be found online at

http://www.lww.com/reprints

Downloaded from circ.ahajournals.org by on May 16, 2011

 Epidemiology and Prevention

Alcohol Intake and Risk of Coronary Heart Disease in

Younger, Middle-Aged, and Older Adults
Ulla A. Hvidtfeldt, MSc; Janne S. Tolstrup, MSc, PhD; Marianne U. Jakobsen, MSc, PhD;
Berit L. Heitmann, PhD; Morten Grnbk, MD, PhD; Eilis OReilly, ScD; Katarina Balter, PhD;
Uri Goldbourt, PhD; Goran Hallmans, MD, PhD; Paul Knekt, PhD; Simin Liu, MD, ScD;
Mark Pereira, PhD; Pirjo Pietinen, ScD; Donna Spiegelman, ScD; June Stevens, MSc, PhD;
Jarmo Virtamo, MD; Walter C. Willett, MD, DrPH; Eric B. Rimm, ScD; Alberto Ascherio, MD, DrPH

BackgroundLight to moderate alcohol consumption is associated with a reduced risk of coronary heart disease. This
protective effect of alcohol, however, may be confined to middle-aged or older individuals. Coronary heart disease
incidence is low in men 40 years of age and in women 50 years of age; for this reason, study cohorts rarely have
the power to investigate the effects of alcohol on coronary heart disease risk in younger adults. This study examined
whether the beneficial effect of alcohol on coronary heart disease depends on age.
Methods and ResultsIn this pooled analysis of 8 prospective studies from North America and Europe including 192 067
women and 74 919 men free of cardiovascular diseases, diabetes, and cancers at baseline, average daily alcohol intake
was assessed at baseline with a food frequency or diet history questionnaire. An inverse association between alcohol and
risk of coronary heart disease was observed in all age groups; hazard ratios among moderately drinking men (5.0 to 29.9
g/d) 39 to 50, 50 to 59, and 60 years of age were 0.58 (95% confidence interval [CI], 0.36 to 0.93), 0.72 (95% CI,
0.60 to 0.86), and 0.85 (95% CI, 0.75 to 0.97) compared with abstainers. However, the analyses indicated a smaller
incidence rate difference between abstainers and moderate consumers in younger adults (incidence rate difference, 45
per 100 000; 90% CI, 8 to 84) than in middle-aged (incidence rate difference, 64 per 100 000; 90% CI, 24 to 102) and
older (incidence rate difference, 89 per 100 000; 90% CI, 44 to 140) adults. Similar results were observed in women.
ConclusionAlcohol is also associated with a decreased risk of coronary heart disease in younger adults; however, the
absolute risk was small compared with middle-aged and older adults. (Circulation. 2010;121:1589-1597.)
Key Words: age groups alcohol consumption coronary disease epidemiology

T he association between alcohol intake and coronary

heart disease (CHD) has been thoroughly investigated
over the past decades with regard to both the amount and
younger adults is limited. Most results are obtained from
cohorts consisting of middle-aged and older adults, and
only a few studies have addressed the effects of alcohol on
type consumed.1 4 In recent years, the importance of CHD in younger adults.9,10 In principle, the cause of CHD
drinking pattern has also been considered.57 In general, among younger adults may differ from that among older
alcohol intake is consistently linked with a lower risk of individuals; for instance, relatively more cases of CHD
CHD. Age-specific incidence rates of CHD vary consid- among younger adults may be attributable to genetic
erably, being very low in men 40 years of age and in causes.11,12 Hence, alcohol may not necessarily protect
women 50 years of age.8 For this reason, the statistical against CHD in this age group. We pooled data from 8
power to investigate the effects of alcohol on CHD in studies to increase sample size and to enable the investi-

Received June 18, 2009; accepted February 12, 2010.

From the Centre for Alcohol Research, National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark (U.A.H., J.S.T.,
M.G.); Social Medicine Section, Department of Public Health, University of Copenhagen, Copenhagen, Denmark (U.A.H.); Research Unit for Dietary
Studies at the Institute of Preventive Medicine, Copenhagen University Hospital, Centre for Health and Society, Copenhagen, Denmark (B.L.H.);
Department of Clinical Epidemiology, Aarhus University Hospital, Aalborg, Denmark (M.U.J.); Departments of Nutrition (E.J.O., W.C.W., E.B.R.,
A.A.), Epidemiology (S.L., D.S., W.C.W., E.B.R., A.A.), and Biostatistics (D.S.), Harvard School of Public Health, Boston Mass; Channing Laboratory,
Department of Medicine, Brigham and Womens Hospital/Harvard Medical School, Boston Mass (W.C.W., E.B.R., A.A.); Department of Medical
Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden (K.B.); Section of Epidemiology and Biostatistics, Henry N. Neufeld Cardiac
Research Institute, Department of Epidemiology and Preventive Medicine, Tel Aviv University, Tel Aviv, Israel (U.G.); Department of Public Health and
Clinical Medicine, Umeaa University, Umeaa, Sweden (G.H.); National Public Health Institute, Helsinki, Finland (P.K., P.P., J.V.); Division of
Epidemiology and Community Health, School of Public Health, University of Minnesota, Minneapolis (M.A.P.); and Departments of Nutrition and
Epidemiology, School of Public Health, University of North Carolina, Chapel Hill (J.S.).
Guest Editor for this article was Wendy S. Post, MD, MS.
Correspondence to Janne S. Tolstrup, PhD, Senior Researcher, Centre for Alcohol Research, National Institute of Public Health, Oster Farimagsgade
5A, 2nd floor, DK-1399 Copenhagen K. E-mail jst@niph.dk
2010 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.109.887513

1589
Downloaded from circ.ahajournals.org by on May 16, 2011
1590 Circulation April 13, 2010

gation of associations between alcohol intake and CHD in after the first follow-up period, and cases were included in whichever
subsets of populations defined by age group. segment they occurred.

Clinical Perspective on p 1597 Measurements

Average daily alcohol intake was assessed at baseline with a food
frequency or diet history questionnaire inquiring about typical intake
Methods of alcoholic beverages. For each beverage, grams of daily alcohol
Study Population intake were calculated from information on the amount and fre-
The analyses were based on data from the Pooling Project of Diet quency of consumption and the alcohol content of the beverage.
and Coronary Disease. The inclusion criteria were the following: a Study-specific conversion factors for the alcohol content were used.
prospective study with at least 150 incident coronary cases, assess- A standard drink contains 10 to 15 g pure alcohol. Total alcohol
ment of usual dietary intake, and a validation study of the diet intake was given by the sum of the beverage-specific intakes.
assessment method. The following 12 studies met these criteria and The outcome of interest was incident CHD events (both fatal and
agreed to share data: Adventists Health Study (AHS),13 Atheroscle- nonfatal). All 8 included studies used validated methods to define
rosis Risk in Communities Study (ARIC),14 -Tocopherol, nonfatal and fatal CHD cases.24
-Carotene Cancer Prevention Study (ATBC),15 Finnish Mobile
Clinic Health Examination (FMC),16 Glostrup Population Study Statistical Methods
(GPS),17 Health Professionals Follow-Up Study (HPFS),18 Israeli Participants were excluded if they reported energy intakes beyond 3 SD
Ischemic Heart Disease Study (IIHD),19 Iowa Womens Health from the study-specific, log-transformed mean energy intake of the
Study (IWHS),20 Nurses Health Study (NHS),21 Vasterbotten Inter- baseline population (1% of the study population). Persons 35 years of
vention Program (VIP),22 and Womens Health Study (WHS).23 The age or with a history of cardiovascular disease, diabetes, or cancers
AHS included only nondrinkers, and the FMC and IIHD studies were (other than nonmelanoma skin cancer) were also excluded. Participants
excluded from the present analysis because of missing information were followed up from baseline to date of CHD event, date of death, or
on alcohol intake. In addition, IWHS was excluded from main end of follow-up, whichever occurred first. Follow-up periods 10
analyses because of self-reported information on CHD. The 8 years (ARIC and GPS) were truncated to reduce heterogeneity. Individ-
remaining studies are presented in Table 1. The NHS was divided ual studies were combined through the use of an aggregated pooled
into 2 segments, thereby taking advantage of repeated assessments analysis technique allowing for calculation of a single exposure-effect
on dietary intake and the long follow-up period. The 2 segments are estimate while adjusting for study origin.25
referred to as NHSa (1980 to 1986) and NHSb (1986 to 1996). The The hazard ratios of CHD were ascertained by the Cox
second segment comprised only women who remained free of CHD proportional-hazards regression model with age as underlying time

Table 1. Baseline Characteristics of Included Studies

CHD Cases, n Alcohol Intake

Study and Baseline Year of Mean Age (90% Follow-Up, CHD Total CHD Median (5th95th Abstainers,
Sex Cohort* Questionnaire Limit), y Person-y Deaths Events Percentile), g/d %
ARIC
Male 5217 19871989 54.6 (4564) 45 652 51 267 12.1 (1.956.6) 45.8
Female 6462 19871989 53.9 (4564) 58 019 18 122 6.8 (1.530.2) 69.8
ATBC
Male 21 141 19841988 57.3 (5069) 121 813 534 1339 13.7 (0.862.3) 10.2
GPS
Male 1658 19741995 51.9 (3580) 14 365 79 102 20.8 (2.772.6) 6.2
Female 1666 19741995 51.5 (3580) 14 605 34 34 8.9 (1.335.1) 18.0
HPFS
Male 41 754 19861988 53.4 (3977) 383 206 421 1273 9.7 (1.046.1) 23.0
NHSa
Female 81 415 19801982 47.1 (3566) 513 915 97 397 5.6 (0.835.0) 31.4
NHSb
Female 61 706 19861988 52.6 (3966) 607 049 208 696 4.9 (0.935.9) 34.2
VIP
Male 9521 19921996 49.1 (3970) 39 230 38 134 4.4 (0.215.8) 4.7
Female 10 555 19921996 49.3 (3970) 43 872 4 23 1.7 (0.17.3) 11.4
WHS
Female 37 272 19921995 53.9 (3889) 190 755 10 152 3.7 (0.928.4) 40.0
Total
Male 79 291 19741996 54.0 (3580) 604 266 1123 3115 9.6 (0.950.4) 18.5
Female 199 076 19741996 50.4 (3589) 1 428 216 371 1424 4.7 (0.835.0) 33.9
*Sample size after exclusion of participants with baseline cardiovascular diseases, cancers, diabetes mellitus, and missing information on alcohol intake.
Median values were calculated for drinkers only.

Downloaded from circ.ahajournals.org by on May 16, 2011

 Hvidtfeldt et al Alcohol and CHD in Age Strata 1591

Table 2. Baseline Characteristics of Women in the Pooled Cohort According to Daily Alcohol Intake
Alcohol Intake, g/d

Total Nondrinkers 0.1 4.9 5.0 14.9 15.0 29.9 30.0 59.9 60.0
n* 192 067 65 121 67 187 39 177 12 258 7418 906
CHD events, n 1365 596 390 241 65 58 15
Age, mean (SD), y 50.4 (7.6) 51.0 (7.7) 49.8 (7.7) 50.1 (7.5) 50.5 (7.4) 51.1 (7.3) 51.3 (7.2)
Education, low, n (%) 8276 (4) 2186 (3) 4719 (7) 936 (2) 322 (3) 96 (1) 17 (2)
Smokers, n (%) 44 513 (23) 11 971 (18) 14 547 (22) 10 454 (27) 3628 (30) 3426 (46) 487 (54)
BMI, mean (SD), kg/m2 25.0 (10.1) 26.0 (5.3) 25.1 (4.5) 24.0 (3.8) 23.7 (3.6) 23.9 (3.8) 24.4 (4.2)
Physical inactivity, n (%) 66 248 (25) 24 382 (37) 21 858 (33) 12 760 (33) 3859 (31) 2971 (40) 418 (46)
Diet, median (5th95th
percentile)
Polyunsaturated fat, g/d 5.4 (3.38.4) 5.5 (3.48.6) 5.4 (3.48.4) 5.4 (3.48.4) 5.3 (3.28.3) 4.8 (2.87.9) 4.0 (2.26.9)
Monounsaturated fat, g/d 13.1 (8.120.5) 13.1 (7.920.7) 13.0 (8.320.5) 13.2 (8.420.5) 12.9 (8.219.7) 12.3 (7.719.0) 10.6 (6.416.8)
Saturated fat, g/d 12.8 (7.720.0) 12.7 (7.520.1) 13.0 (8.020.2) 12.9 (7.920.1) 12.6 (7.619.7) 11.9 (7.218.6) 10.3 (5.716.8)
Fiber, g/d 15.4 (8.425.7) 15.7 (8.526.7) 15.9 (8.926.1) 15.0 (8.324.4) 14.2 (7.823.1) 12.2 (6.720.4) 10.3 (5.018.9)
Cholesterol, mg/d 254 (139458) 252 (134464) 249 (137451) 264 (148460) 263 (151464) 252 (143443) 222 (121394)
Total energy, kcal/d 1603 (9002627) 1579 (8712637) 1581 (8932589) 1614 (9152608) 1678 (9772689) 1758 (10702749) 2008 (12833050)
Hypertension, n (%) 35 637 (19) 13 566 (21) 11 695 (17) 6271 (16) 2137 (17) 1695 (23) 273 (30)
Dyslipidemia, n 20 850 (11) 8389 (13) 6599 (10) 3787 (10) 1217 (10) 731 (10) 127 (14)
Differences in distribution of covariates across levels of alcohol consumption were tested by ANOVA (age, BMI), Kruskal-Wallis (dietary factors), and 2 tests
(education, smoking, physical activity, hypertension, and dyslipidemia). All tests showed statistically significant differences (P0.0001).
*After exclusion of participants with missing information on any of the relevant covariates.
Defined as less than high school.
Energy adjusted.

scale, allowing for delayed entry (left censoring).26 Absolute risks score of weekly time spent on various activities during the past year
(incidence rates) describing the scale of CHD according to sex, age, (ARIC, HPFS, NHS, VIP, and WHS) or according to the intensity of
and level of alcohol intake were estimated by means of Poisson the average weekly physical activity during the past 12 months
regression.27 Absolute risk differences were calculated, and 90% (ATBC, GPS). These measures were harmonized to a 5-level
confidence limits were derived by bootstrap estimation (5000 repli- variable from 1 (least active) to 5 (most active).30 32 In addition,
cations) with the 5th and 95th percentiles of the distribution as the models were stratified by study origin and baseline year to account
lower and upper limits. for differences in follow-up procedures or questionnaire design and
We performed primary analyses considering the risk of CHD in period effects. Information on postmenopausal hormone therapy use
categories of alcohol consumption (0, 0.1 to 4.9, 5.0 to 14.9, 15.0 to was unavailable in VIP; therefore, the main analyses for women did
29.9, 30.0 to 59.9, and 60.0 g/d in women; 0, 0.1 to 4.9, 5.0 to 14.9, not include adjustment for this factor. Sensitivity analyses included
15.0 to 29.9, 30.0 to 59.9, 60.0 to 89.9, and 90.0 g/d in men) both measures of self-reported history of physician-diagnosed elevated
for each individual study and for the pooled cohort. The study cholesterol (dyslipidemia) and hypertension (yes/no).
population was analyzed separately in the following 3 age groups: 39
to 49.9, 50 to 59.9, and 60 years. Age was updated during Results
follow-up, and participants were assigned to the appropriate age
category; thus, each person could contribute person-time at risk to Baseline Characteristics
1 age category. Baseline characteristics of participants from the 8 included
Additional analyses exploring the risk of CHD per alcohol studies are shown in Table 1. In total, the pooled study
increment (1 g/d) were performed. Alcohol was modeled continu-
ously using second-degree fractional polynomials, thus allowing for population comprised 199 076 women and 79 291 men who
a single turning point (the nadir) in the risk function. Following the experienced 1424 and 3115 coronary events during 1 428 216
results of Corrao and others,1 a model describing the dose-response and 604 266 person-years of follow-up, respectively. Base-
relationship of alcohol on CHD including both a linear and root- line age of participants varied from 35 to 89 years in women
squared term of alcohol was applied. and from 35 to 80 years in men with a mean age of 50.4 and
The P value for the test for trend was obtained by assigning the
median value within categories of alcohol intake and using this 54.0 years, respectively. The proportion of nondrinkers varied
variable as a continuous variable. SAS statistical package version 9.1 substantially between studies from 11.4% to 53.0% in women
was used for all analyses.28 and from 4.7% to 45.8% in men. Median alcohol intakes
We harmonized the variables of the different studies, and the varied from 4.4 to 20.8 g/d in men and from 1.7 to 8.9 g/d in
following set of potential confounders was identified on the basis of
women.
the method of causal diagrams as suggested by Greenland and
others29: educational level (less than high school, high school, more Tables 2 and 3 show characteristics of participants in-
than high school), smoking (never smokers, ex-smokers, and current cluded in the pooled cohort according to alcohol consump-
smokers of 1 to 4, 5 to 14, 15 to 24, or 25 cigarettes per day), body tion. Heavier alcohol intake was associated with higher
mass index (BMI; 18.5, 18.5 to 24.9, 25.0 to 29.9, and 30 proportions of smokers, physical inactivity, and hypertension
kg/m2), total energy intake (kcal/d), and energy-adjusted quintiles of
cholesterol, dietary fiber, saturated fat, monounsaturated fat, and and lower median intakes of fat and fiber; a moderate alcohol
polyunsaturated fat intake. Physical activity measures varied across intake was associated with the highest median cholesterol
the cohorts, measured either according to an energy expenditure intake compared with abstainers and heavy drinkers. In men,
Downloaded from circ.ahajournals.org by on May 16, 2011
1592 Circulation April 13, 2010

Table 3. Baseline Characteristics of Men in the Pooled Cohort According to Daily Alcohol Intake
Alcohol Intake, g/d

Total Nondrinkers 0.1 4.9 5.0 14.9 15.0 29.9 30.0 59.9 60.0 89.9 90.0

n* 74 720 13 904 19 030 20 556 11 375 7789 1546 520 (1)

CHD events, n 2961 623 737 751 449 311 59 31
Age, mean (SD), y 54.0 (8.4) 54.4 (8.6) 53.3 (8.7) 53.6 (8.5) 54.6 (7.6) 55.2 (7.6) 55.2 (7.6) 55.8 (6.2)
Education, low, n (%) 20 773 (28) 2513 (18) 6125 (32) 5580 (27) 3590 (32) 2283 (29) 450 (29) 232 (45)
Smokers, n (%) 28 144 (38) 3338 (24) 6086 (32) 7409 (36) 5688 (50) 4254 (55) 952 (62) 417 (80)
BMI, mean (SD), 25.8 (8.2) 26.0 (3.8) 25.8 (3.4) 25.7 (3.4) 25.8 (3.4) 25.9 (3.5) 26.1 (3.9) 26.5 (4.1)
kg/m2
Physical inactivity, 18 850 (25) 3740 (27) 4411 (23) 4489 (22) 2956 (26) 2421 (31) 572 (37) 261 (80)
n (%)
Diet, median
(5th95th
percentile)
Polyunsaturated 5.4 (3.38.9) 5.5 (3.39.0) 5.3 (3.48.8) 5.4 (3.48.9) 5.4 (3.29.0) 5.2 (3.08.8) 4.6 (2.68.2) 4.1 (2.27.8)
fat, g/d
Monounsaturated 12.8 (8.516.8) 12.9 (8.017.3) 12.8 (8.616.8) 12.9 (8.816.8) 12.9 (9.016.6) 12.4 (8.316.3) 11.3 (7.415.4) 10.5 (6.815.5)
fat, g/d
Saturated fat, g/d 13.2 (7.523.8) 12.3 (6.922.8) 13.6 (7.724.0) 13.3 (7.724.0) 13.9 (7.824.5) 13.2 (7.423.5) 12.7 (6.521.6) 13.3 (6.521.0)
Fiber, g/d 19.8 (11.532.3) 20.7 (11.635.7) 21.0 (12.733.5) 20.1 (12.331.8) 19.1 (11.530.1) 17.0 (10.127.4) 14.6 (8.324.5) 12.9 (6.621.2)
Cholesterol, mg/d 322 (174573) 315 (161568) 303 (166536) 322 (177566) 349 (193602) 343 (190610) 323 (176606) 324 (167544)
Total energy, kcal/d 2135 (11493671) 1929 (10443438) 2047 (11203537) 2111 (11733598) 2311 (12743823) 2387 (13443911) 2652 (15184138) 3054 (15704662)
Hypertension, n (%) 14 388 (19) 2759 (20) 3292 (17) 3634 (18) 2246 (20) 1902 (24) 421 (27) 134 (26)
Dyslipidemia, n 5076 (9) 1145 (10) 1114 (8) 1312 (9) 720 (11) 609 (13) 147 (17) 29 (16)

Differences in distribution of covariates across levels of alcohol consumption were tested by ANOVA (age, BMI), Kruskal-Wallis (dietary factors), and 2 tests
(education, smoking, physical activity, hypertension, and dyslipidemia). All tests showed statistically significant differences (P0.0001).
*After exclusion of participants with missing information on any of the relevant covariates.
Defined as less than high school.
Energy-adjusted.
This information was not available for ATBC.

a higher frequency low educational level was observed and pooled estimates for women and men. In women, an
among participants in the highest alcohol group. Similar inverse relation between alcohol intake and CHD was
distributions of covariates according to alcohol intake were found in each individual study except for VIP. The
observed across cohorts. confidence bounds around risk estimates for this particular
study were very broad because the reference category
Alcohol Consumption and Risk of CHD included only 2 cases. In men, an inverse relation was
Tables 4 and 5 show the relative risks (hazard ratios) of observed in all studies. In the pooled analysis, we observed
CHD by categories of alcohol intake for each of the studies a significantly lower risk of CHD among women with an

Table 4. Study-Specific and Pooled Hazard Ratios of CHD for Categories of Daily Alcohol Intake for Women
Hazard Ratio (95% CI) by Daily Alcohol Intake, g/d
Hazard Ratio,
Nondrinkers 0.1 4.9 5.0 14.9 15.0 29.9 30.0 59.9 60.0 P for
n (CHD606) (CHD397) (CHD242) (CHD65) (CHD60) (CHD15) Trend
Study specific
ARIC 6406 1.00 (Reference) 0.84 (0.441.59) 0.58 (0.291.17) 0.78 (0.302.00) NA NA 0.0738
GPS 1509 1.00 (Reference) 0.75 (0.272.06) 0.89 (0.312.55) 1.15 (0.314.23) 0.52 (0.055.21) NA 0.8418
NHSa 79 479 1.00 (Reference) 0.73 (0.570.94) 0.72 (0.550.95) 0.59 (0.380.94) 0.49 (0.290.82) 1.64 (0.773.49) 0.1548
NHSb 60 083 1.00 (Reference) 0.78 (0.650.94) 0.67 (0.540.84) 0.47 (0.320.69) 0.57 (0.390.82) 0.60 (0.261.38) 0.0003
VIP 9758 1.00 (Reference) 2.26 (0.2818.47) 2.17 (0.1238.67) NA NA NA 0.8654
WHS 34 832 1.00 (Reference) 1.02 (0.691.50) 0.81 (0.491.35) 0.37 (0.111.19) 0.58 (0.181.90) 1.32 (0.189.95) 0.1768
Pooled
Age adjusted* 192 067 1.00 (Reference) 0.78 (0.690.89) 0.73 (0.620.84) 0.57 (0.440.74) 0.80 (0.611.05) 1.71 (1.022.86) 0.1112
Smoking adjusted 192 067 1.00 (Reference) 0.75 (0.650.85) 0.62 (0.530.72) 0.47 (0.360.61) 0.52 (0.390.68) 1.02 (0.611.70) 0.0001
Multivariable 192 067 1.00 (Reference) 0.78 (0.690.90) 0.68 (0.590.80) 0.52 (0.400.67) 0.53 (0.390.70) 0.93 (0.551.58) 0.0001
*Also adjusted for year of baseline questionnaire.
Also adjusted for age and year of baseline questionnaire.
Multivariable hazard ratios were adjusted for age, year of baseline questionnaire, smoking, BMI, education, physical activity, energy intake, polyunsaturated fat,
monounsaturated fat, saturated fat, fiber, cholesterol intake, and study origin.
Not applicable because of a limited number of cases.

Downloaded from circ.ahajournals.org by on May 16, 2011

 Hvidtfeldt et al Alcohol and CHD in Age Strata 1593

Table 5. StudySpecific and Pooled Hazard Ratios of CHD for Categories of Daily Alcohol Intake for Men
Alcohol Intake, g/d
Hazard Ratio,
Nondrinkers 0.1 4.9 5.0 14.9 15.0 29.9 30.0 59.9 60.0 89.9 90.0 P for
n (CHD623) (CHD737) (CHD751) (CHD449) (CHD311) (CHD59) (CHD31) Trend
Study specific
ARIC 5166 1.00 (Reference) 1.16 (0.781.71) 1.29 (0.931.79) 0.87 (0.571.32) 0.73 (0.441.22) 0.74 (0.272.04) 1.40 (0.563.54) 0.4499
ATBC 21 119 1.00 (Reference) 0.86 (0.711.03) 0.86 (0.721.03) 0.70 (0.580.85) 0.61 (0.490.77) 0.48 (0.320.74) 0.80 (0.511.27) 0.0001
GPS 1294 1.00 (Reference) 0.84 (0.302.32) 0.54 (0.211.38) 0.60 (0.241.50) 0.44 (0.161.18) 0.37 (0.091.52) NA 0.0304
HPFS 38 654 1.00 (Reference) 1.00 (0.851.17) 0.75 (0.630.88) 0.69 (0.560.86) 0.66 (0.520.83) 0.65 (0.411.01) 0.17 (0.021.23) 0.0001
VIP 8486 1.00 (Reference) 0.50 (0.280.93) 0.24 (0.110.48) 0.25 (0.070.91) 0.73 (0.095.85) NA NA 0.009
Pooled
Age-adjusted* 74 719 1.00 (Reference) 0.95 (0.851.06) 0.84 (0.750.93) 0.75 (0.660.85) 0.74 (0.640.85) 0.70 (0.530.91) 0.96 (0.671.39) 0.0001
Smoking-adjusted 74 719 1.00 (Reference) 0.94 (0.841.05) 0.81 (0.720.90) 0.71 (0.630.81) 0.66 (0.580.76) 0.60 (0.460.79) 0.81 (0.571.18) 0.0001
Multivariable 74 719 1.00 (Reference) 0.96 (0.861.08) 0.83 (0.740.92) 0.72 (0.640.82) 0.66 (0.570.76) 0.58 (0.440.77) 0.77 (0.531.13) 0.0001
*Also adjusted for year of baseline questionnaire.
Also adjusted for age and year of baseline questionnaire.
Multivariable hazard ratios were adjusted for age, year of baseline questionnaire, smoking, BMI, education, physical activity, energy intake, polyunsaturated fat,
monounsaturated fat, saturated fat, fiber, cholesterol intake, and study origin.
Not applicable because of a limited number of cases.

alcohol intake of up to 60 g/d and among men with an Incidence rates of CHD in women and men according to
alcohol intake of up to 90 g/d. alcohol intake and age are displayed in Figure 3. As expected,
We also performed analyses describing the risk of CHD the incidence of CHD was much lower in the younger
according to alcohol consumption modeled as a continuous compared with older participants. The incidence rates of
variable (Figure 1). In both men and women, a reduction in CHD among female abstainers in the 3 age groups were 11
CHD risk was observed at low to moderate levels of alcohol. (95% CI, 1 to 109), 41 (95% CI, 1 to 400), and 103 (95% CI,
The relative risk of CHD was 0.58 (95% confidence interval 9 to 1018) per 100 000, respectively. In male abstainers, the
[CI], 0.49 to 0.68) in women and 0.69 (95% CI, 0.62 to 0.76) incidence rates were 114 (95% CI, 77 to 171), 262 (95% CI,
in men with a daily intake of 30 g/d, corresponding to 2 to 201 to 343), and 454 (95% CI, 354 to 553) per 100 000 for the
3 drinks. Higher levels of alcohol consumption were not 3 age groups, respectively. In all age groups and in both men
and women, the incidence rate was lower among participants
associated with any discernible additional protection in
with a low to moderate alcohol intake compared with abstain-
women and with only modest protection in men.
ers. In women, the incidence rate differences between drink-
ing 0 g/d and 5.0 to 29.9 g/d were 3 (90% CI, 1 to 25), 16
Alcohol Consumption and Risk of CHD in
(90% CI, 0 to 111), and 35 (90% CI, 0 to 250). For men, the
Age Strata
corresponding incidence rate differences were 45 (90% CI, 8
We estimated the risks of CHD according to alcohol intake
to 84), 64 (90% CI, 24 to 102), and 89 (90% CI, 44 to 140)
separately for 3 age groups (50, 50 to 59, and 60 years;
per 100 000.
Figure 2). In all age groups and for both men and women, a
decreased risk of CHD according to alcohol intake was Sensitivity Analyses
observed but with broader confidence bounds for the young- Heterogeneity between study-specific effects was assessed by
est age group. The test for interaction between alcohol and the inclusion of an interaction term between alcohol and
age was not statistically significant in either women (P0.34) study origin under the null hypothesis of no between-study
or men (P0.25). We also modeled the risk continuously for differences in the relative risk of CHD by alcohol intake,1,25
the different age groups and observed similar shapes of the with no sign of heterogeneity detected (P0.95 in women,
curves (data not shown). P0.12 in men). In addition, comparing pooled risk estimates

Figure 1. Relative risk functions (95%

CI) describing the dose-response rela-
tion between alcohol intake and risk of
CHD. Analyses were adjusted for year of
baseline questionnaire, education, smok-
ing, BMI, physical activity, total energy
intake, polyunsaturated fat, monounsatu-
rated fat, saturated fat, fiber, and choles-
terol intake. The fitted model (SE) is
reported. The 99th percentile of cases
was 64 g/d in women and 90 g/d in
men. The highest alcohol intake in
observed cases was 90 g/d in women
and 215 g/d in men.

Downloaded from circ.ahajournals.org by on May 16, 2011

1594 Circulation April 13, 2010

Figure 2. Sex- and age-specific pooled

hazard ratios of CHD for categories of
daily alcohol intake. Multivariable hazard
ratios were adjusted for year of baseline
questionnaire, education, smoking, BMI,
physical activity, total energy intake,
polyunsaturated fat, monounsaturated
fat, saturated fat, fiber, and cholesterol
intake. n indicates number cases for
each sex and age group.

after systematically excluding each study at a time confirmed Additional analyses for women were performed to examine
that no single study strongly influenced the pooled esti- whether adjustment for postmenopausal hormone replace-
mates.33 Hence, the pooled hazard ratios are considered ment therapy had an impact on results. This involved exclud-
appropriate summaries of the study-specific data. Performing ing participants with unavailable information on this partic-
a test for interaction between age (time scale) and alcohol ular covariate (n9799, 5% of the study population). In the
consumption did not yield violations of the proportional- remaining cohorts, postmenopausal hormone replacement
hazards assumption for either women (P0.10) or men therapy did not appreciably affect the association between
(P0.22). alcohol and risk of CHD. In addition, the inclusion of history
Separate analyses were performed for fatal and nonfatal of hypertension and dyslipidemia as covariates in the model
events to examine whether the effect of alcohol on CHD did not affect the risk estimates of CHD according to alcohol
differed according to the severity of the outcome. The results consumption. Furthermore, an analysis was performed in-
of this analysis did not reveal obvious differences between cluding the IWHS (n29 801). This inclusion also did not
the 2 measures of outcome, although a tendency toward an change the hazard ratios appreciably (data not shown).
elevated risk of fatal CHD was observed for the highest A test for interaction between alcohol and smoking was
alcohol category in both women and men (data not shown). performed to examine whether the estimates of the effect
To examine the possibility that latent baseline symptoms of of alcohol on CHD risk differed according to smoking
CHD might reduce the alcohol intake, thereby biasing the status. Smoking did not modify the association between
results, we performed analyses in which the first 2 or 4 years alcohol and CHD in either men (P0.79) or women
after baseline were excluded. This did not attenuate the (P0.14). Additional analyses including nonsmokers only
estimates (data not shown). were performed separately for women (n100 144) and
Downloaded from circ.ahajournals.org by on May 16, 2011
 Hvidtfeldt et al Alcohol and CHD in Age Strata 1595

alcohol intake and diabetes mellitus could also explain some

of the benefit from alcohol intake.35,36 In addition, alcohol has
an effect on plasminogen activator inhibitor-1 that would tend
to reduce thrombosis.37
Previous studies have suggested that the causes of CHD in
younger adults differ from the mechanisms involved with
CHD in older persons. Results from the Honolulu Heart
Program indicated that the effect of hypertension, BMI, and
cholesterol on CHD differed according to age. For instance,
the relative risk of CHD in hypertensive men declined from
3.7 in those 45 to 54 years of age to 1.7 in those 75 years
of age. Similarly, associations between BMI and total cho-
lesterol weakened with advancing age.10 The Coronary Artery
Risk in Young Adults (CARDIA) study of men and women
33 to 45 years of age found that alcohol intake was associated
with expected dose response between alcohol and high-
density lipoprotein cholesterol levels and an inverse relation
between alcohol and fibrinogen levels.9 They also observed
an increased risk of coronary calcification with greater
alcohol consumption. Because coronary calcification is a
marker of atherosclerosis, this result in young adults is not
consistent with the results of the present study.9
Another aspect of alcohol consumption related to age is
drinking patterns. Younger adults may tend to binge drink
more often than older persons, which may increase their risk
of CHD5,7,9; however, findings from the CARDIA study did
Figure 3. Incidence rates of CHD according to age and alcohol not indicate a protective effect of alcohol intake on coronary
intake. Analyses were adjusted for year of baseline question- calcification in younger adults even after the exclusion of
naire, education, smoking, BMI, physical activity, total energy binge drinkers.9
intake, polyunsaturated fat, monounsaturated fat, saturated fat,
fiber, and cholesterol intake.
Our findings suggest a J-shaped curve in women, but in
men, the risk did not increase significantly at high amounts of
alcohol. Biomarkers that mediate the association between
men (n46 576) for alcohol levels of 0, 0.1 to 14.9, 15.0 alcohol and decreased risk of CHD such as high-density
to 29.9, and 30.0 g/d and showed similar associations lipoprotein and fibrinogen are found to explain a larger
(data not shown). proportion of the association among men than among women,
which may indicate that alcohol has specific effects on such
Discussion mediators according to sex.7 Other biological explanations for
In this pooled cohort of 8 prospective studies, we observed a sex-specific associations include differences in alcohol phar-
lower risk of CHD among men and women with a light to macokinetics (ie, processing and elimination of alcohol in the
moderate alcohol intake compared with nondrinkers. This body), which depend largely on body composition.38,39 How-
finding was consistent across different age groups without ever, because the risk curves of this study were modeled
significant variations in dose response. separately for the 2 sexes, comparisons between them are not
The current data on the effect of alcohol on CHD in straightforward.
younger adults are sparse. In a study based on the Honolulu The present study is one of only a few existing studies
Heart Program, the authors compared CHD risk according to focusing on the effects of alcohol on CHD according to age.
conventional risk factors in middle-aged and older men (age, Our work was based on a large body of data with thorough
45 to 93 years). Compared with nondrinkers, they observed a measurements of alcohol intake and relevant covariates. A
lower risk of CHD among drinkers in middle-aged but not strength was the availability of diet data in the Pooling
among older participants (75 years) and concluded accord- Project of Diet and Coronary Disease that enabled adjustment
ingly that the relation between alcohol and CHD weakened for potential dietary confounders. The size of the study
with age. The study, however, was limited by the simple population allowed us to perform subset analyses exploring
categorization of alcohol intake into drinkers or nondrinkers the association between alcohol and CHD in strata of younger
and included men only.10 men and women, aspects that even large individual cohorts do
Several plausible explanations for the lowered risk of CHD not have the power to address. The findings of the present
among moderate drinkers exist. Among those explanations, study were strengthened by the prospective design, which
the evidence is probably strongest for a mechanism involving provided information on the sequence of events allowing for
alcohol increasing high-density lipoprotein cholesterol and conclusions on causality, assuming proper confounding con-
reducing plasma fibrinogen levels, thereby reducing platelet trol. Potential confounders of the association between alcohol
aggregability.7,34 The hypothesized J-shaped relation between and CHD were carefully selected on the basis of directed
Downloaded from circ.ahajournals.org by on May 16, 2011
1596 Circulation April 13, 2010

acyclic graphs, ensuring a minimally sufficient set of covari- was not collected in the database of the Pooling Project of
ates. Furthermore, the inclusion criteria of the Pooling Project Diet and Coronary Disease.
of Diet and Coronary Disease enabled adjustment for relevant
dietary factors, which most previous studies did not control Conclusions
for. The pooled analyses included both cohorts and interven- This study supports current knowledge that alcohol in mod-
tion studies from North America and Europe, and similar erate amounts protects against CHD in both men and women.
effects were observed across the studies. Finally, an advan- Our findings further suggest that this effect is also present in
tage of the Pooling Project is the inclusion of previously younger age groups. However, younger adults are at low risk
unpublished results, thereby reducing the risk of publication for CHD, and the beneficial effects obtained by a moderate
bias. alcohol intake may be negligible compared with the increased
However, several limitations of the study should also be risk of, for instance, traffic accidents and cancer. Recommen-
considered. We focused on the importance of the amount of dations on alcohol intake among younger adults should
alcohol consumed; however, other aspects of patterns of consider all-cause mortality and morbidity.
alcohol intake may be equally important and were not
addressed. Our study included only information on current Sources of Funding
alcohol consumption and confounders at baseline. For this This research was supported by research grant R01 HL58904 from
reason, the reference category of abstainers may contain the National Heart, Lung, and Blood Institute of the National
former drinkers who quit because of existing illness, which Institutes of Health. The Unit for Dietary Studies was funded by the
Female Researchers in Joint Action program of the Danish Medical
could cause a moderate alcohol intake to appear more Research Council.
protective than it is. Although several studies that included
only lifelong or long-term abstainers in this category have Disclosures
confirmed a protective effect of alcohol even among healthy None.
individuals,6,40 the sick-quitter hypothesis is relevant in the
present context because older age groups may include more References
abstainers who stopped drinking because of illness (eg, 1. Corrao G, Rubbiati L, Bagnardi V, Zambon A, Poikolainen K. Alcohol
hypertension). and coronary heart disease: a meta-analysis. Addiction. 2000;95:
15051523.
As mentioned, patterns of alcohol consumption (eg, choice 2. Grnbk M, Becker U, Johansen D, Gottschau A, Schnohr P, Hein HO,
of alcohol type and frequency of consumption) may differ Jensen G, Srensen TI. Type of alcohol consumed and mortality from all
considerably with age, which we did not account for. Our causes, coronary heart disease, and cancer. Ann Intern Med. 2000;133:
411 419.
findings of protective effects of alcohol on CHD in all
3. Rimm EB, Giovannucci EL, Willett WC, Colditz GA, Ascherio A,
examined age groups may indicate that neither the type of Rosner B, Stampfer MJ. Prospective study of alcohol consumption and
alcohol nor the frequency of consumption modifies the risk of coronary disease in men. Lancet. 1991;338:464 468.
influence of alcohol on CHD considerably. However, future 4. Fuchs CS, Stampfer MJ, Colditz GA, Giovannucci EL, Manson JE,
Kawachi I, Hunter DJ, Hankinson SE, Hennekens CH, Rosner Bl.
research based on observational studies should emphasize
Alcohol consumption and mortality among women. N Engl J Med. 1995;
other measures of alcohol intake such as frequency of alcohol 332:12451250.
consumed. Furthermore, cohort studies with information on 5. Tolstrup J, Jensen MK, Tjnneland A, Overvad K, Mukamal KJ,
lifetime alcohol intake or repeated measurements of alcohol Grnbk M. Prospective study of alcohol drinking patterns and coronary
heart disease in women and men. BMJ. 2006;332:1244 1248.
intake and potential confounders could contribute valuable 6. Mukamal KJ, Conigrave KM, Mittleman MA, Camargo CA Jr, Stampfer
insight because changes in alcohol intake over a given period MJ, Willett WC, Rimm EB. Roles of drinking pattern and type of alcohol
may be of great importance. Additional experimental studies consumed in coronary heart disease in men. N Engl J Med. 2003;348:
are also needed to expand the knowledge of biological 109 118.
7. Mukamal KJ, Jensen MK, Grnbk M, Stampfer MJ, Manson JE,
mechanisms. Pischon T, Rimm EB. Drinking frequency, mediating biomarkers, and
Furthermore, this study focused only on CHD events. risk of myocardial infarction in women and men. Circulation. 2005;112:
Overall effects of alcohol on all-cause morbidity and mortal- 1406 1413.
ity must be considered if we are to optimize alcohol guide- 8. Stamler J. Established major risk factors: historical overview. In: Marmot
M, Elliott P, eds. Coronary Heart Disease Epidemiology: From Aetiology
lines for different age groups of the population. The lower to Public Health. New York, NY: Oxford University Press; 2005:18 31.
risk of all-cause mortality is expected to be caused mainly by 9. Pletcher MJ, Varosy P, Kiefe CI, Lewis CE, Sidney S, Hulley SB.
the effects of alcohol on CHD. In this study, a lower risk of Alcohol consumption, binge drinking, and early coronary calcification:
CHD was observed in all examined age groups in moderate findings from the Coronary Artery Risk Development in Young Adults
(CARDIA) Study. Am J Epidemiol. 2005;161:423 433.
alcohol consumers compared with abstainers; however, the 10. Abbott RD, Curb JD, Rodriguez BL, Masaki KH, Yano K, Schatz IJ,
absolute risk of CHD was rather low in the youngest age Ross GW, Petrovitch H. Age-related changes in risk factor effects on the
group. Thus, considering the increasing contribution of CHD incidence of coronary heart disease. Ann Epidemiol. 2002;12:173181.
11. Marenberg ME, Risch N, Berkman LF, Floderus B, de Faire U. Genetic
to all-cause mortality with age, it is reasonable to assume that
susceptibility to death from coronary heart disease in a study of twins.
the protective effect of alcohol on all-cause mortality is N Engl J Med. 1994;330:10411046.
mostly pronounced in older age groups. This issue has been 12. Zdravkovic S, Wienke A, Pedersen NL, Marenberg ME, Yashin AI, de
addressed in a few previous studies indicating that the Faire U. Heritability of death from coronary heart disease: a 36-year
follow-up of 20 966 Swedish twins. J Intern Med. 2002;252:247254.
protective effect of alcohol consumption on mortality in
13. Fraser GE, Sabate J, Beeson WL, Strahan TM. A possible protective
general is confined to middle-aged and older individu- effect of nut consumption on risk of coronary heart disease: the Adventist
als.4,41,42 Unfortunately, information on all-cause mortality Health Study. Arch Intern Med. 1992;152:1416 1424.

Downloaded from circ.ahajournals.org by on May 16, 2011

 Hvidtfeldt et al Alcohol and CHD in Age Strata 1597

14. ARIC Investigators. The Atherosclerosis Risk in Communities (ARIC) 27. Woodward M. Modelling follow-up data. In: Woodward M, ed. Epide-
Study: design and objectives. Am J Epidemiol. 1989;129:687702. miology: Study Design and Data Analysis. Boca Raton, Fla: Chapman &
15. ATBC Cancer Prevention Study Group. The Alpha-Tocopherol, Beta- Hall; 2005:601 671.
Carotene Lung Cancer Prevention Study: design, methods, participant 28. SAS/STAT Software [computer program]. Version 9.1 for Windows. Cary,
characteristics, and compliance. Ann Epidemiol. 1994;4:110. NC: SAS Institute Inc; 2002.
16. Knekt P, Reunanen A, Jarvinen R, Seppanen R, Heliovaara M, Aromaa 29. Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic
A. Antioxidant vitamin intake and coronary mortality in a longitudinal research. Epidemiology. 1999;10:37 48.
population study. Am J Epidemiol. 1994;139:1180 1189. 30. Devore EE, Kang JH, Okereke O, Grodstein F. Physical activity levels
17. Hoyer AP, Engholm G. Serum lipids and breast cancer risk: a cohort and cognition in women with type 2 diabetes. Am J Epidemiol. 2009;
study of 5,207 Danish women. Cancer Causes Control. 1992;3:403 408. 170:1040 1047.
31. Tanasescu M, Leitzmann MF, Rimm EB, Hu FB. Physical activity in
18. Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett
relation to cardiovascular disease and total mortality among men with
WC. Reproducibility and validity of an expanded self-administered semi-
type 2 diabetes. Circulation. 2003;107:24352439.
quantitative food frequency questionnaire among male health profes-
32. Hemila H, Kaprio J, Albanes D, Virtamo J. Physical activity and the risk
sionals. Am J Epidemiol. 1992;135:1114 1126.
of pneumonia in male smokers administered vitamin E and beta-carotene.
19. Balogh M, Medalie JH, Smith H, Groen JJ. The development of a dietary Int J Sports Med. 2006;27:336 341.
questionnaire for an ischemic heart disease survey. Isr J Med Sci. 1968; 33. Morton LM, Hartge P, Holford TR, Holly EA, Chiu BC, Vineis P,
4:195203. Stagnaro E, Willett EV, Franceschi S, La Vecchia C, Hughes AM, Cozen
20. Munger RG, Folsom AR, Kushi LH, Kaye SA, Sellers TA. Dietary W, Davis S, Severson RK, Bernstein L, Mayne ST, Dee FR, Cerhan JR,
assessment of older Iowa women with a food frequency questionnaire: Zheng T. Cigarette smoking and risk of non-Hodgkin lymphoma: a
nutrient intake, reproducibility, and comparison with 24-hour dietary pooled analysis from the International Lymphoma Epidemiology Con-
recall interviews. Am J Epidemiol. 1992;136:192200. sortium (InterLymph). Cancer Epidemiol Biomarkers Prev. 2005;14:
21. Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, 925933.
Hennekens CH, Speizer FE. Reproducibility and validity of a semiquan- 34. Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. Moderate
titative food frequency questionnaire. Am J Epidemiol. 1985;122:51 65. alcohol intake and lower risk of coronary heart disease: meta-analysis of
22. Hallmans G, Agren A, Johansson G, Johansson A, Stegmayr B, Jansson effects on lipids and haemostatic factors. BMJ. 1999;319:15231528.
JH, Lindahl B, Rolandsson O, Soderberg S, Nilsson M, Johansson I, 35. Klatsky AL. Alcohol, cardiovascular diseases and diabetes mellitus.
Weinehall L. Cardiovascular disease and diabetes in the Northern Sweden Pharmacol Res. 2007;55:237247.
Health and Disease Study Cohort: evaluation of risk factors and their 36. Koppes LL, Dekker JM, Hendriks HF, Bouter LM, Heine RJ. Moderate
interactions. Scand J Public Health Suppl. 2003;61:18 24. alcohol consumption lowers the risk of type 2 diabetes: a meta-analysis of
23. Rexrode KM, Lee IM, Cook NR, Hennekens CH, Buring JE. Baseline prospective observational studies. Diabetes Care. 2005;28:719 725.
characteristics of participants in the Womens Health Study. J Womens 37. Lee KW, Lip GY. Effects of lifestyle on hemostasis, fibrinolysis, and
Health Gend Based Med. 2000;9:19 27. platelet reactivity: a systematic review. Arch Intern Med. 2003;163:
24. Pereira MA, OReilly E, Augustsson K, Fraser GE, Goldbourt U, 2368 2392.
Heitmann BL, Hallmans G, Knekt P, Liu S, Pietinen P, Spiegelman D, 38. Makela P, Paljarvi T, Poikolainen K. Heavy and nonheavy drinking
occasions, all-cause and cardiovascular mortality and hospitalizations: a
Stevens J, Virtamo J, Willett WC, Ascherio A. Dietary fiber and risk of
follow-up study in a population with a low consumption level. J Stud
coronary heart disease: a pooled analysis of cohort studies. Arch Intern
Alcohol. 2005;66:722728.
Med. 2004;164:370 376.
39. Mumenthaler MS, Taylor JL, OHara R, Yesavage JA. Gender dif-
25. Smith-Warner SA, Spiegelman D, Ritz J, Albanes D, Beeson WL, ferences in moderate drinking effects. Alcohol Res Health. 1999;23:
Bernstein L, Berrino F, van den Brandt PA, Buring JE, Cho E, Colditz 55 64.
GA, Folsom AR, Freudenheim JL, Giovannucci E, Goldbohm RA, 40. Klatsky AL, Armstrong MA, Friedman GD. Risk of cardiovascular mor-
Graham S, Harnack L, Horn-Ross PL, Krogh V, Leitzmann MF, tality in alcohol drinkers, ex-drinkers and nondrinkers. Am J Cardiol.
McCullough ML, Miller AB, Rodriguez C, Rohan TE, Schatzkin A, 1990;66:12371242.
Shore R, Virtanen M, Willett WC, Wolk A, Zeleniuch-Jacquotte A, 41. White IR, Altmann DR, Nanchahal K. Alcohol consumption and mor-
Zhang SM, Hunter DJ. Methods for pooling results of epidemiologic tality: modelling risks for men and women at different ages. BMJ. 2002;
studies: the Pooling Project of Prospective Studies of Diet and Cancer. 325:191.
Am J Epidemiol. 2006;163:10531064. 42. Gmel G, Gutjahr E, Rehm J. How stable is the risk curve between alcohol
26. Clayton D, Hills M. Coxs Method for Follow-Up Studies: Statistical and all-cause mortality and what factors influence the shape? A precision-
Models in Epidemiology. New York, NY: Oxford; 1993:298 306. weighted hierarchical meta-analysis. Eur J Epidemiol. 2003;18:631 642.

CLINICAL PERSPECTIVE
The association between alcohol consumption and decreased risk of coronary heart disease is well established, but possible
age differences are plausible because of potential pathogenic differences in coronary heart disease events occurring in
younger compared with middle-aged or older individuals. We studied these relations in a large population of men and
women 35 to 89 years of age at baseline. We observed a lower risk of coronary heart disease among men and women with
a light to moderate alcohol intake compared with nondrinkers, and this finding was consistent and of similar size in all age
groups. However, the absolute risk of CHD was small in the youngest age group, and risk differences between abstainers
and light to moderate alcohol consumers were of negligible size. Therefore, our results provide strong evidence for a lower
risk of coronary heart disease among moderate consumers relative to nondrinkers in younger, middle-aged, and older
adults; however, considering absolute risks across age groups, younger adults are not likely to benefit from an overall
recommendation of moderate alcohol intake.

Downloaded from circ.ahajournals.org by on May 16, 2011