Received 4/1995: revised 7/17/95; accepted 7/18/95.

I Supported by USPIIS Grant CA17054 from the National Cancer Institute, NIH,Department of Health and Human Services, and by the California Public Health

Foundation. subcontract 050.F-87t)9, which is supported by the California Dc-

partment of Health Services as part of its statewide cancer-reporting program

mandated by Health and Safety Code Sections 21(1 and 21 1.3. The ideas and

opinions expressed herein are those of the authors. and no endorsement by the

state of California. I)epartment of Health Services. or the California Public Health

Foundation is intended or should he inferred.

2 To whom requests for reprints should be addressed. at Epidemiology Branch,

National Institute (If Environmental Health Services, P. 0. B)X 12233. MD

A3-05. Research Triangle Park, NC 27709.

Vol. 4. 72 1-725. October/November /995 Cancer Epidemiology. Biomarkers & Prevention 72/

Lifetime Alcohol Consumption and Breast Cancer Risk among

Postmenopausal Women in Los An’

Matthew P. Longnecker,2 Annlia Paganini-Hill, andRonald K. Ross

Department (If EpidemioIog�. University of California at Los Angeles School

of Public Health, Los Angeles. California 90095-1772 [M. P. LI, Department

of Preventive Medicine. University of Southern (‘alifornia School of Medicine

lM. P. L.. A. P-El.. R. K. RI, and University of Southem Califomia Norris

Comprehensive (‘ancer (‘enter ER. K. RI, Los Angeles. California 901)33

Abstract

Although most data support an overall relation of alcohol

consumption with risk of breast cancer, the timing ofalcohol consumption as it relates to risk of breast canceris still debatable. The authors examined this issue in a

case-control study conducted among non-Hispanic whiteand Hispanic women in Los Angeles. Cases aged 55-64years at diagnosis in 1987-1989 were enrolled throughthe Cancer Surveillance Program of Los Angeles County(a Surveillance, Epidemiology, and End Results Programregistry). Community controls were individually matchedto cases by age (± 3 years), ethnicity, and neighborhood.In-person interviews were completed with 1510 matched

pairs, of which 1431 met the inclusion criteria for thepresent analysis. In a multivariate conditional logisticregression model that simultaneously included terms foralcohol consumption at age 25 years, age 40 years, and inthe recent past, clear differences among the age-specificassociations were not observed, and the results suggestedthat intake at each time independently contributed torisk. The odds ratios associated with estimated averagelifetime intake were: for <6 g/day, 1.01; for 6-11 g/day,1.21; for 12-18 g/day, 0.94; for 19-32 g/day, 1.63; for 33-

45 g/day, 2.45 and for � 46 g/day, 0.94 compared withabstainers (P for trend = 0.01). Use of estrogenreplacement therapy did not modify the risk associatedwith alcohol consumption, in contrast with the findings intwo previous studies. This large study supports a relationbetween risk of breast cancer and alcohol consumptionand suggests that lifetime intake may be the bestindicator of alcohol-associated risk.

Introduction

Most epidemiological studies that have addressed breast cancerrisk and alcohol use have shown that the risk of breast cancer

is greater among women who consume alcoholic beverages ( I).The increase in risk, however, is subtle: study results varywidely, and the mechanism for this association has yet to heelucidated. This has left uncertainty about whether alcoholcauses breast cancer and about the level of risk associated witha given alcohol intake. Furthermore, some studies (2-5) purportthat consumption before age 30 years is more strongly associ-

ated with risk than is consumption at later ages, but not allstudies support this risk pattern (6-9). Because these issueshave important implications for women consuming alcohol or

contemplating alcohol use and in understanding the biologicalbasis for this relation, we investigated breast cancer risk andlevels of alcohol consumption at different ages by using datafrom a large population-based case-control study in Lk)S Ange-les County, CA.

Materials and Methods

Selection of Cases and Controls. Eligible subjects (cases andcontrols) were female, English-speaking, non-Hispanic whitesor Hispanic, residents of Los Angeles County, and born in the

United States, Canada, or western Europe.Eligible cases were identified through the Cancer Surveil-

lance Program of Los Angeles County, a population-based

Surveillance, Epidemiology, and End Results cancer registrycovering the more than nine million residents of Los AngelesCounty. Patients were aged 55-64 years at the time of a firstdiagnosis of histologically confirmed breast cancer betweenMarch 1, 1987 and December 31, 1989. Both invasive and insitu breast cancers were included. During the accrual period,2497 eligible cases were identified. Of these cases, 9 1 1 werenot interviewed for the following reasons: physician recom-mended against contact (129); subject refused interview (417),was too ill, or dead (229); had moved outside the county (1 13):

or we were unable to locate the subject (23). We interviewed

the remaining 1586 cases or 64% of those eligible. Theseincluded 1425 patients with invasive and 161 with in sitit

disease.Eligible controls were individually matched to cases by

age (± 3 years), ethnicity, and neighborhood. We interviewedI control for each of 1510 cases and were unable to interviewan eligible control for the remaining 76 cases. Potential controlswere identified by canvasing housing units in the neighborhoodwhere the case subject lived at the time of diagnosis by using

a predefined walk pattern. For each unit selected, we tried todetermine whether an eligible matched control lived in thehousehold. If no one was home at the time of the visit, we maderepeated efforts to contact the inhabitants by mail and phone.The number of housing units contacted to identify an eligiblecontrol was distributed with quartiles 9, 25, and 59; e.g. , 50%of controls were identified after contacting 25 units. For 1208

on March 4, 2020. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

722 Lifetime Alcohol Consumption and Breast Cancer Risk

3 The abbreviations used are: OR. odds ratio; Cl, confidence interval.

cases, the first eligible control participated. One eligible personrefused before the recruitment of a matched control for 228

cases: 2 refused for 65 cases: 3 refused for 8 cases: and 4refused for 1 case. Among controls identified as potentially

eligible, the response rate was 80% { lSlO/[1208 + 2(228) +

3(65) + 4(8) + 5(1)]}. While attempting to find controls for the76 unmatched cases, we contacted a median of 105 housing

units and encountered 25 refusals. Complete censuses wereobtained in the walk patterns for neighborhoods of 649 cases.

At some addresses, we were denied access to an apartmentbuilding or were otherwise unable to complete the census. The

median number of unknowns in the walk patterns of the re-

maining cases was four. The study was limited to non-Hispanicwhite and Hispanic women primarily because this method of

control ascertainment has proved highly inefficient for subjectsof other ethnicities in similarly designed studies.

The study protocol was approved by the University of

Southern California School of Medicine Institutional Review

Board.

Data Collection. The subjects were interviewed in their homesby trained study staff. The interview took on average about 45

mm to complete. The same staff member generally interviewedboth members of each case-control pair.

The questions on alcohol use addressed average consump-tion at three times in the subjects’ lives: age 25 years, age 40

years, and at the reference age (1 year before the case’s diag-nosis or, for controls, I year before the corresponding case’s

diagnosis). Usual weekly consumption of beer, wine, and liquorwas ascertained separately. The instrument also included de-

tailed questions on hormone use, established breast cancer risk

factors. and other characteristics.

Data Analysis. Of the 1510 matched sets, 53 contained 1subject reporting menstrual periods in the reference year with-

out corresponding use of hormone replacement therapy. Onesubject in an additional 26 sets had missing data for alcohol or

a covariate (listed below). This analysis was restricted to post-

menopausal women with complete alcohol and covariate data

and, therefore, included 1431 case-control sets.For women whose age at menopause was unknown and

could not he inferred from date of total ovariectomy, a value of50 years was assigned (the median age for those with a reported

age at menopause).We assumed one serving of beer contained 12.8 g of

alcohol, OflC 4-ounce glass of wine, 10.9 g, and one mixeddrink, 15.0 g (10). Categories of alcohol consumption weredetermined ci priori (9) and were chosen so that the medianintake within the higher categories corresponded approximately

to I (12-18 g alcohol), 2 (19-32 g), 3 (33-45 g), or 4 (�46 g)

alcoholic drinks daily.Lifetime alcohol consumption was estimated as the aver-

age alcohol consumption for ages 25 and 40 years, and recentintake.

Multivariate conditional regression models were fit byusing EGRET (1 1). Results from univariate models were notsubstantially different. The covariates included were as follows:

age at menarche (�Il, 12, 13, and �14 years): education (lessthan high school, high school, some college, or college ormore): benign breast disease (no or yes): family history ofbreast cancer (io or yes); body mass index (<21.9, 22.0-24.0,

24.1-27.4, and �27.5 kg/m2): parity (ordinal), age at first

full-term pregnancy (nulliparous, <20, 20-29, and �30 years);

age at menopause (<45, 45-49, 50-54, and �55 years); and

ethnicity (noit-Hispanic white and Hispanic). As a test of trend(two-sided P) we fit a coefficient to alcohol intake as an ordinal

Table I Number of subjects and adjusted ORs (and 95% CIs) for breast

cancer according to alcohol consumption at age 25 years, age 40 years. and inthe recent past, with and without mutual adjustment for intake at other times.

Results are based on 1431 matched sets.

Alcohol intake.

(g/day) Cases

No. (If

Controls

Adjusted OR”

.Mutually Adjusted

- � �

No Yes

Intake at age 25 yr

(1 934 978 1

>0-S 256 222 1.20 (0.97-1.52) 1.19 (0.94-1.52)

6-1 I I 17 122 0.96 (0.72-1.28) 0.95 (0.69-1.30)

I 2-18 57 51 1 . 17 (0.78-1.77) 1 .04 (0.66-1.62)

19-32 38 37 1.1 1 (0.67-1.83) 0.90(0.53-1.53)

33-45 16 8 2.17 (0.89-5.29) 1.84 (0.73-4.63)

�46 13 13 0.99 (0.44-2.20) 0.90 (0.38-2.15)

trend P 0.25 0.84

Intake at age 40 yr

t) 685 713 1

>()-5 265 284 1.00(0.81-1.23) 0.99 (0.77-1.28)

6-1 1 189 186 1.01 (0.79-1.28) 1.05(0.78-1.42)

12-18 117 103 1.21 (0.89-1.65) 1.08 (0.75-1.57)

19-32 96 87 1.28 (0.92-1.80) 1.15 (0.77-1.73)

33-45 35 20 2.32 (1.27-4.25) 1.89 (0.95-3.75)

�46 44 38 1.1 1 (0.7(1-1.77) t).99 (0.57-1.71)

trend P 0.1)3 0.42

Intake in recent past

(1 774 773 1

>0-5 204 228 0.90(0.71-1.14) 0.88 (0.67-1.15)

6-I I 131 173 0.73 (0.55-4)96) 0.70(0.51-0.94)

12-18 123 103 1.31 (0.96-1.79) 1.20(0.84-1.69)

19-32 1 15 97 1.28 (0.93-1.76) 1.14 (0.79-1.64)

33-45 47 3)) 1.56 (0.94-2.59) 1.21 (0.69-2.13)

�46 37 27 1.36(0.79-2.35) 1.24(0.68-2.26)

trend P 0.02 0.35

“ Adjusted for age at menarche, education, benign breast disease, family history,body mass index, parity, age at first full-term pregnancy. age at menopause. and

ethnicity (see “Materials and Methods”). Mutually adjusted means, e.g., that the

result for drinking at age 25 years has been adjusted for intake at age 40 years and

recent intake.

variable. Tests for interaction were performed by comparing theslope for lifetime alcohol intake as a continuous variable across

levels of potential modifiers.

Results

Table 1 provides data on breast cancer risk and in relation toestimated alcohol consumption at ages 25 and 40 years and in

the recent past, with and without adjustment for intake at othertimes. The results without adjustment for intake at other timesshowed that recent consumption of I 2 or more g of alcohol (oneor more alcoholic drinks) daily was associated with a small

increase in risk of breast cancer, although the ORs3 wereimprecise. The test for trend, however, supported an increasedrisk in women who recently drank alcohol. When the relation ofrecent intake with risk was examined after adjusting for intake

at other times, the association was attenuated. Attenuation ofassociations after adjusting for intake at other times was alsoevident in the results for ages 25 and 40 years. The results forages 25 and 40 years, adjusted for intake at other times, showedrelative risks of 1.8-1.9 associated with consumption of 33-45

on March 4, 2020. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

Table 2 Number of subjects and adjusted ORs (and 95% CIs) for breast

cancer according to estimated lifetime average alcohol consumption (average

of age 25 years. age 4(1 years. and recent past). Results are based on 1431

matched sets.

Benign breast disease

No

Yes

Menanche

�11

12

13

�14

0.91) t).74 t).89(0.69-l.l4)

1.51 1.79 1.16(0.93-1.46)

l.t)2 1.20 I.08(t).95-l.24)

t).87 1.80 I.60”(1.I0-2.32)

Age at menopause

<45

45-49

5(1-54

55 +

I 1.22 1.74’ 1.17 (t).97-l.40)

I t).93 1.40 1.23 (0.99-1.53)

I l.()6 1.05 1.01 (0.81-1.26)

I I . I 2 0.83 05)#{176})(0.81)-I .21)

Education

�1Ith grade

High school

Some college

College or more

Ethnicity

non-Hispanic white

Hispanic

1.29 1.1)9(0.96-1.24)

1.17 1.1 1 (t).94-I.3I)

1.2(1 1.1)8 (11.97-1.21)

1.32 l.l7(t).9I-1.51)

1 t).93 1.08 t).97 ((1.67-1.42)

I 1.07 1.26 1.11(1(10-1.23)

Cancer Epidemiology, Biomarkers & Prevention 723

Alcohol intake

(glday) Cases

No. of

Controls OR

Adjusted”

(95% CI)

()

>0-5

6.-Il

12-18

19-32

33-45

�46

534

427

222

101

101

3(1

16

561

453

201

116

7t)

13

17

1

1.01

1.21

0.94

1.63

2.45

0.94

(0.84-1.22)

(0.95-1.55)

(0.69-1.29)

(1.14-2.33)

(1.22-4.93)

(0.46-1.93)

trend P 0.01

“ Adjusted for age at menarche, education, benign breast disease, family history,body mass index, parity, age at first full-term pregnancy, age at menopause, and

ethnicity (see “Materials and Methods”).

g alcohol/day (about 3 drinks) but no increase in risk amongwomen reporting �46 g/day (4 or more drinks).

Women whose estimated lifetime average alcohol intakewas 19-45 g (2-3 drinks) daily experienced a 63-145% in-

crease in risk of breast cancer, and the test for trend supported

an overall association (Table 2). Women consuming on average�46 g/day (4 or more drinks/day), however, had a relative risk

near one.We examined whether the risk associated with alcohol

consumption varied by the type of alcoholic beverage. Contin-

uous terms for the lifetime average amounts of alcohol con-sumed in the different beverage types were entered simulta-neously in a multivariate model (same covariates as in Table 1).The relative risks associated with 13 g/day (about one drink)were: beer, 0.91 (95% CI 0.71-1.17); wine, 1.04 (95% CI0.79-1.36), and spirits, 1.22 (95% CI 1.04-1.42). The riskassociated with spirit intake did not differ significantly from

that for other beverages (likelihood ratio test, � = 3.21;degrees of freedom, 1).

To determine if a change in alcoholic beverage consump-tion after age 40 years was associated with breast cancer risk,we fit a multivariate-adjusted model with a continuous term foralcohol consumption at age 40 years and a second continuousterm representing the change between consumption at age 40years and recent consumption. The OR per 13 g/day for con-sumption at age 40 years was 1.14 (95% CI 1.04-1.24), and

the OR per 13 g/day change from the amount at age 40 was 1.12

(95% CI = 1.03-1 .23). The model can be used to calculate thata woman who had 1 drink daily at age 40 years who subse-quently quit drinking would have an OR for breast cancer of0.89 (95% CI = 0.81-0.97) compared with a woman who hadcontinued drinking that amount.

We examined whether the association between estimatedaverage lifetime alcohol intake and breast cancer risk varied

according to level of other established or suggested breastcancer risk factors and whether the disease was invasive or in

situ (Table 3). The association of alcohol intake with breast

cancer was stronger among w.:tnen without a family history ofbreast cancer than it was among those with such a history but

the difference was not statistically significant (P > 0.05). Apersonal history of benign breast disease did not modify riskassociated with alcohol intake. The alcohol-breast cancer asso-ciation tended to be larger among women with menarche beforeage 13 years: in another model (data not shown) we fit a.:( .‘ficient for the interaction of alcohol consumption and age

Table 3 Adjusted ORs for breast cancer according to estimated

average lifetime alcohol intake, stratified by level of other factors(n = 1431 pairs).”

Factor level --�

Lifetime�- � ----

alcohol�---..- � �.

intake (gJday)-- � .---�-�- - -

0 >9-18 �I9 Per 13 glday

(All women) I 1.06 1.24 l.lO(I.(X)-l.21)”

Family history (1st degree)

No I 1.11 1.29’ I.I2(l.t)I-l.25)

Yes I 0.81 0.97 0.99(0.77-1.28)

Age at first full-term pregnancy

Nulliparous I

<2() 1

20-29 1

30+ 1

1 l.t)8 1.21 .1(1(0.99-1.23)

I 1.1$) 1.34 1.09 (t).86- 1.38)

I 1.06 1.28 1.18 (t).94-I.47)

1 0.97 1.68’ 1.34 (1.08-1.65)

1 1.26 1.09 0.96(0.78-1.18)

1 0.95 1.01) 1 .1)1 (t).84- I .2 1

Parity

0 1 0.90 1)74 0.89 (0.69-1.14)

1-2 1 1.03 1.34 1.21 (1.1)3-1.41)

3-4 1 1.15 1.43 1.15(0.97-1.36)

5+ 1 t).96 1.1 1 0.98(0.73-1.31)

Body mass index (kg/m2)

<21.9

22.0-24.0

24.1-27.4

�27.5

Ever used unopposed estrogen

1 t).93 1.36 1.16 (0.89-1.52)

1 0.94 t).76 0.97 (t).80-1.I8)

1 1.14 1.57’ 1.19(1.03-1.37)

I 0.94 0.46 11.75 (0.44-1.26)

I 11.89 1.52 1.12 (0.85-1.47)

1 0.89 0.94 1 .03 (0.88-I .2 1

1 1.34 1.34 l.t)8 (0.90-1.30)

I 1.08 1.48 1.28(1.00-1.63)

I 1.04 1.22 l.lt)(0.99-I.22)

1 1.36 1.72 1.24 (0.70-2.19)

No I 1.1)6

Yes I I .05

Ever used estrogen and progesterone�

No 1 1.11)

Yes I 0.85

Ever used oral contraceptive

No I 1(8) 1.42’ 1.22(1.08-1.37)

t).94 0.7(1 0.73�(0.57-4l93)

Degree of invasiveness

In situ

Invasive

“ Adjusted for ag� at mer,�rche, eJu’ation, benign breast disease, faril, hstory.body mass index, parity. age at first full-term pregnancy, age at menopause. and

ethnicity (see “Materials and Methods”).

S � CI in parentheses.

‘- 95% CI excludes one.

d p < 0.05 for difference from nulliparous group.

P At any time.

fP < 0.001 for difference from never-use group.

Yes

on March 4, 2020. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

724 Lifetime Alcohol Consumption and Breast Cancer Risk

at menarche as continuous values, and its P value was 0.08. Thestrongest risk modification with alcohol consumption amongthe subgroups examined was age at first full-term pregnancy;women who were nulliparous experienced no increased risk

with alcohol consumption, whereas in women whose first full-term pregnancy was at age 30 years or later, the association wasrelatively strong (OR 1.60/13 g/day). We repeated this anal-ysis to examine the interaction of drinking at age 25 years with

age at first full-term pregnancy and found results similar tothose shown. The alcohol-breast cancer association tended to be

larger among women with a body mass index below 24.1kg/rn2: in another model (data not shown), we fit a coefficientfor the interaction of alcohol consumption and body mass index

as continuous values, and its P value was 0.15. The alcohol-breast cancer association was greater among Hispanics than

among non-Hispanic whites, but the difference was not statis-tically significant. Use of estrogen replacement therapy, con-sidered in various ways including long-term or current use (datanot shown), did not alter the alcohol-breast cancer association;this was also true when the relations were examined by usingrecent alcohol intake. Alcohol consumption was associatedwith a reduced risk of breast cancer among women who had

ever used oral contraceptives (OR = 0.73; 95% CI = 0.57-

0.93); the same pattern was evident when drinking at youngerages (average of intake at 25 and 40 years) was considered.

Alcohol was a stronger risk factor for invasive as comparedwith in situ breast cancer, but the difference in associations wasnot statistically significant.

Discussion

This large study supports an association of alcohol consump-lion with increased risk of breast cancer. Our findings regardingalcohol intake in early adulthood, although not definitive, wereconsistent with a contribution to risk not markedly greater thanthat associated with later consumption. Furthermore, in thisstudy of women aged 55-64 years, the data suggest that de-

creasing alcohol consumption after age 40 years can reduce riskof breast cancer.

When we examined the alcohol-breast cancer association

as most others have, i.e. , by considering the relation of recentintake to risk unadjusted for intake at other times in life (TableI ), the magnitude of the association was weak and almostexactly the same as that found in a meta-analysis of 38 studieson alcohol and breast cancer (1). When we adjusted recentintake for intake at other times, however, the association was

attenuated, suggesting that intake in other periods also contrib-uted to risk. In fact, the strongest association of alcohol and riskof breast cancer was observed when lifetime consumption was

considered, similar to the result in another recent study (9). Inthe present study, lifetime consumption of 19-45 g/day wasclearly associated with increased risk. Lifetime consumption of�46 g alcohol/day (4 or more drinks), however, was associated

with a relative risk of breast cancer of 0.94. This odd finding

may be due to a complex pattern of selection bias, misreportingof alcohol consumption among heavy drinkers, or confounding.Alternatively, because premenopausal alcoholic women areknown to have an excess prevalence of amenorrhea (12), the

associated reduction in exposure to endogenous estrogens orprogestin could account for this finding. Another recent largestudy also found that women drinking �46 �/day were at lowerrisk that those drinking 33-45 g’day (9).

Our findings are consistent with the evidence that drinking

in early adulthood is not especially associated with risk ofbreast cancer (6-9). Most data to the contrary (2-5) are not

particularly strong or convincing. With the present data, how-ever, we could not assess the association of drinking before age25 years with risk.

If early-age alcohol consumption is reported with mark-edly less precision than recent consumption, then this differ-ence in measurement error could mask a stronger association ofearly age drinking with risk. However, Dwyer et a!. (13) foundthat alcohol consumption recorded at age 30 years correlatedwell (r = 0.78) with consumption recalled for age 30 years

when subjects were studied again at age 50 years. Other data

also support that alcohol consumption reported for periods in

the remote past is as reliable as reports of more recent con-sumption (14).

In these data, consumption of alcohol in distilled bever-

ages was more strongly associated with risk of breast cancerthan was consumption in beer or wine; the difference amongbeverages, however, was not statistically significant. In otherstudies, no specific beverage has been consistently implicatedover others (1, 9). Martin-Moreno et a!. (15) have noted thetendency of American studies to find that wine consumption isnot a risk factor, whereas the converse is true in most Europeanstudies.

Our data showed that the alcohol-breast cancer associationwas greater in women whose age at first birth was at age 30

years or more, and that the association was protective amongwomen who had ever used oral contraceptives. Although datain two other studies also showed the alcohol-breast cancerassociation to be slightly stronger among women with a late age

at first birth (16, 17), other data are not particularly supportiveof a modifying effect ( 18, 19). In other data (6, 1 7), variation in

risk by oral contraceptive use was not apparent. Previous re-

ports suggesting that the alcohol-breast cancer association isgreater among users of replacement estrogens (1 7, 20) were not

confirmed by our data nor by other recent studies (21, 22). A

greater alcohol-breast cancer association in leaner women, ob-served in two early reports (19, 23), was weakly supported by

our study and others (17, 24) but has been absent elsewhere (6,25). Other results on interaction with age at menarche show nosystematic pattern (16, 17). In summary, consistent evidencethat any of the factors considered modify the alcohol-breast

cancer association is not present.The response rates in the present study were less than

ideal, and the possibility of recall bias cannot be entirely ex-eluded. Of controls identified as potentially eligible, 80% par-ticipated. The response rate among all eligible controls, al-though not calculable because our census was incomplete, wassomewhat lower. Although only 64% of all eligible cases wereinterviewed, less than one-half of those not interviewed (17%

of the total) were lost due to a reason plausibly related toalcohol consumption (subject refusal). In two studies (26, 27)

conducted among women with breast cancer whose prediag-nostic alcohol intake had been ascertained both before and after

diagnosis, recall of past consumption appeared not to be ma-terially altered by the diagnosis. The magnitude of the alcohol-breast cancer association in our data was comparable to thatobserved in follow-up studies (1), in which bias due to selectionand recall was less likely or nonexistent.

As in any study with data based on self-reported alcoholuse, the degree of bias due to misreporting of alcohol consump-

tion is unknown. Data presented by Giovannucci et a!. (28),

however, suggest that among nonalcoholic populations, thereported amount of recent alcohol intake (ascertained withquestions like those in the present study) correlates highly with

consumption noted in diet records.Some unknown factor that is strongly related to alcohol

on March 4, 2020. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

Cancer Epidemiology, Biomarkers & Prevention 725

consumption and is also a risk factor for breast cancer couldaccount for the association of alcohol with breast cancer, but the

existence of such a factor seems improbable. A complete dis-cussion of whether alcohol causes breast cancer (1, 29) isbeyond the scope of this paper; these data do, however, addweight to the evidence that an association exists.

If alcohol consumption causes breast cancer, its actionmay be mediated through an effect on estrogen metabolism (30,

31). Given the multiplicity of proposed mechanisms by whichalcohol may cause cancer (32, 33), the lack of data identifyingparticular mechanisms as relevant in humans, and the incon-

sistent results of animal models (29), the need for additional

experimental work is clear.In conclusion, these data support the association of alcohol

consumption, especially average lifetime consumption, with

increased risk of breast cancer. Furthermore, the data suggestthat decreasing alcohol consumption in later adulthood mayreduce risk of breast cancer. The potential benefit of decreasing

alcohol consumption, if any, however, needs to be assessedconsidering all potential effects of alcohol consumption (33-35) and a woman’s overall risk profile.

References

I . Longnecker. M. P. Alcohol consumption in relation to risk of breast cancer:

meta-analysis and review. Cancer Causes & Control, 5: 73-82, 1994.

2. Harvey. E. B., Schairer, M. S., Bninton, L. A,, Hoover, R. N.. and Fraumeni,

J. F., Jr. Alcohol consumption and breast cancer. J. NatI. Cancer Inst., 78:

657-661, 1987.

3. Hiatt, R. A., Klatsky. A. L., and Armstrong, M. A. Alcohol consumption and

the risk of breast cancer in a prepaid health plan. Cancer Res., 48: 2284-2287,

1988.

4. Young. T. B. A case-control study of breast cancer and alcohol consumption

habits. Cancer (Phila.). 64: 552-558, 1989.

5. Van’t Veer, P., Kok. F. J., Hermus, R. J. J., and Sturmans, F. Alcohol dose,

frequency and age at first exposure in relation to the risk of breast cancer. Int. J.

Epidemiol.. /8: 511-517, 1989.

6. La Vecchia, C., Negni. E., Parazzini, F., Boyle, P., Fasoli, M., Gentile, A., and

Fnanceschi, S. Alcohol and breast cancer: update from an Italian case-control

study. Eur. j. Cancer Clin. Oncol., 25: 171 1-1717, 1989.

7. Nasca, P. C., Baptiste, M. S.. Field, N. A., Metzger, B. B., Black, M., Kwon,C. S., and Jacobson, H. An epidemiological case-control study of breast cancer

and alcohol consumption. Int. J. Epidemiol., /9: 532-538, 1990.

8. Katsouyanni, K.. Tnichopoulou, A., Stuver, S., Vassilaros, S., Papadiamantis,Y., Bournas. N., Skarpou, N.. Mueller, N., and Tnichopoulos, D. Ethanol andbreast cancer: an association that may be both confounded and causal. Int. J.

Cancer, 58: 356-361, 1994.

9. Longnecker, M. P., Newcomb, P. A., Mittendorf, R., Greenberg, E. R., Clapp,R. W., Bogdan, G. F., Baron, J., MacMahon, B., and Willett, W. C. Risk of breast

cancer in relation to lifetime alcohol consumption. J. NatI. Cancer Inst., 87:

923-929, 1995.

10. United States Department of Agriculture. Composition of foods: beverages;

raw, processed, prepared. Agriculture Handbook No. 8-14. United States Dc-

partment of Agriculture, 1986.

1 1. Statistics and Epidemiology Research Corporation. EGRET: Epidemiologi-

cal Graphics, Estimation, and Testing Package. Seattle, WA: SERC.

12. Mello, N. Effects of alcohol abuse on reproductive function in women.

Recent Dcv. Alcohol., 6: 253-276, 1988.

13. Dwyer. J. T.. Gardner, J., Halvorsen, K., Krall, E. A,, Cohen, A., andValadian, I. Memory of food intake in the distant past. Am. J. Epidemiol.. 130:

1033-1046, 1989.

14. Longnecker, M. P., Newcomb, P. A., Mittendorf, R., Greenberg, E. R., Clapp,

R. W., Bogdan, G., Willett, W. C., and MacMahon, B. The reliability of self-reported alcohol consumption in the remote past. Epidemiology, 3: 535-539,1992.

15. Martin-Moreno, J, M., Boyle P., Gorgojo, L., Willett, W. C., Gonzalez, J.,Villar, F., and Maisonneuve, P. Alcoholic beverage consumption and risk of

breast cancer in Spain. Cancer Causes & Control, 4: 345-353, 1993.

16. van den Brandt, P., Goldbohm, R. A., and van’t Veer, P. Alcohol and breast

cancer: results from the Netherlands cohort study. Am. J. Epidemiol., /41:907-915, 1995.

17. Gapstur, S. M., Potter, J. D., Sellers, T. A., and Folsom, A, R. Increased risk

of breast cancer with alcohol consumption in postmenopausal women. Am. J.Epidemiol., /36: 1221-1231, 1992.

18. La Vecchia, C,, Decarli, A., Franceschi, S., Pampallona, 5,, and Tognoni, G.

Alcohol consumption and the risk of breast cancer in women. J. NatI. Cancer Inst.,

75: 61-65, 1985.

19. Willett, W. C., Stampfer, M. J., Golditz, G. A., Rosner, B. A., Hennekens, C.

H., and Speizer, F. E. Moderate alcohol consumption and the risk of breast cancer.

N. EngI. J. Med., 316: 1 174-1 180, 1987.

20. Colditz, G. A.. Stampfer, M. J.. Willett, W. C., Hennekens, C. H., Rosner, B.,

and Speizer, F. E. Prospective study of estrogen replacement therapy and risk of

breast cancer in postmenopausal women. JAMA, 264: 2648-2653, I99().

21. Newcomb, P. A., Longnecker, M. P., Storer, B. E., Mittendorf, R., Green-berg. E. R., Clapp, R. W., Bogdan, G., and Willett, W. C. Long-term hormonereplacement use and risk of breast cancer in postmenopausal women. Am. J.

Epidemiol., in press, 1995.

22. Fniedenreich, C. M. Re: ‘Increased risk of breast cancer with alcohol con-

sumption in postmenopausal women.” Am. J. Epidemiol.. 139: 541-542, 1994.

23. Schatzkin, A., Jones, D. Y., Hoover, R. N., Taylor, P. R., Brinton, L. A.,

Ziegler, R. G., Harvey, E. B., Carter, C. L,, Licitra, L. M., and Larson, D. B.Alcohol consumption and breast cancer in the epidemiologic follow-up study of

the First National Health and Nutrition Examination Survey. N. EngI. J. Med.,316: 1169-1173, 1987.

24. Hams, R. E., and Wynder, E. L, Breast cancer and alcohol consumption. Astudy in weak associations. JAMA, 259: 2867-2871, 1988.

25. Rosenberg, L., Palmer, J. R., Miller, D. R., Clarke, E. A., and Shapiro, S. A

case-control study of alcoholic beverage consumption and breast cancer. Am. J.Epidemiol., 131: 6-14, 199().

26. Giovannucci E., Stampfer, M. J., Colditz, G. A., Manson, J. E., Rosner, B. A.,

Longnecker, M. P., Speizer, F. E., and Willett, W. C. Recall and selection bias inreporting past alcohol consumption among breast cancer cases. Cancer Causes &

Control, 4: 441-443, 1993.

27. Fniedenreich, C. M., Howe, G. R., and Miller, A. B. An investigation of recall

bias in the reporting of past food intake among breast cancer cases and controls.

Ann. Epidemiol., 1: 439-453, 1991.

28. Giovannucci, E., Colditz, G., Stampfer, M. J., Rimm, E. B., Litin, L.,

Sampson, L., and Willett, W, C. The assessment of alcohol consumption by asimple self-administered questionnaire. Am. J. Epidemiol., /33: 810-817, 1991.

29. Schatzkin, A., and Longnecker, M. P. Alcohol and breast cancer: where arewe now and where do we go from here? Cancer (Phila.), 74: 1 101-1 1 10, 1994.

30. Longnecker, M. P. Do hormones link alcohol with breast cancer? J. NatI.

Cancer Inst., 85: 692-693, 1993.

31. Hankinson, S. E., Willett, W. C., Manson, J. E., Hunter, D. J., Colditz, G. A.,

Stampfer, M. J., Longcope, C., and Speizer, F. E. Alcohol, height, and adiposity

in relation to plasma estrogen and prolactin levels in postmenopausal women. J.

NatI. Cancer Inst., in press, 1995.

32. Garro, A. J., and Lieber, C. S. Alcohol and cancer. Ann. Rev. Pharmacol.

Toxicol., 30: 219-249, 1991).

33, Longnecker, M. P. Alcohol consumption and risk of cancer in humans: an

overview. Alcohol, 12: 87-96, 1995.

34, Maclure, M. Demonstration of deductive meta-analysis: ethanol intake andrisk of myocardial infarction. Epidemiol. Rev., /5: 328-351, 1993.

35. Hurley. J., Horowitz, J., (eds.). Alcohol and Health. New York: HemispherePublishing Corporation. 1990.

on March 4, 2020. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

1995;4:721-725. Cancer Epidemiol Biomarkers Prev   M P Longnecker, A Paganini-Hill and R K Ross  postmenopausal women in Los Angeles.Lifetime alcohol consumption and breast cancer risk among

  Updated version

  http://cebp.aacrjournals.org/content/4/7/721

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cebp.aacrjournals.org/content/4/7/721To request permission to re-use all or part of this article, use this link

on March 4, 2020. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from