Known breast-cancer risk factors account for only part of the variabil
ity in breast-cancer incidence. Tobacco smoke is not commonly consider
ed a breast carcinogen, but many of its constituents, such as N-nitros
amines, are carcinogenic in laboratory animal studies. Herein, we asse
ssed a cytochrome P4502E1 (CYP2E1) genetic polymorphism (a Dral restri
ction enzyme site in intron 6) as a risk factor for breast cancer in b
oth premenopausal and postmenopausal women. Because N-nitrosamines are
metabolically activated by CYP2E1, the risk among women smokers was i
nvestigated. Caucasian women were enrolled in a case-control study of
breast cancer between 1986 and 1991. A subset of the women (219 premen
opausal and 387 postmenopausal women) consented to phlebotomy. The all
elic frequencies for the premenopausal women (D allele = 0.91 and C al
lele = 0.09) and postmenopausal women (D allele = 0.93 and C allele =
0.07) were similar to those previously reported. There was no statisti
cally significant association between the CYP2E1 polymorphism and brea
st-cancer risk for premenopausal or postmenopausal women (adjusted odd
s ratio (OR) = 1.04, 95% confidence interval (CI) = 0.48, 2.24, and OR
= 1.01, 95% CI = 0.55, 1.84, respectively). When the women were categ
orized as nonsmokers versus smokers (those who smoked more than one ci
garette per week for more than 1 yr), premenopausal women with one or
two C alleles who had a history of smoking were found to be at increas
ed risk (unadjusted OR = 7.00, 95% CI = 0.75, 14.53, and adjusted OR =
11.09, 95% Cl = 1.51, 81.41), although the number of study subjects w
ith those genotypes was small. The small number of study subjects with
a C allele precluded meaningful classification by level of smoking, b
ut categorizing the smokers into two groups (above and below the media
n) also suggested an increased risk. Premenopausal women with the DD g
enotype and postmenopausal women with any genotype were not at increas
ed risk. Breast-cancer risk was not related to the CYP2E1 genotype in
either premenopausal nonsmokers or smokers (adjusted OR = 0.66, 95% CI
= 0.20, 2.17, and OR = 2.13, 95% CI = 0.60, 7.59, respectively) or po
stmenopausal nonsmokers or smokers (OR = 0.90, 95% CI = 0.34, 2.35, an
d OR = 1.02, 95% CI = 0.46, 2.23, respectively), although the differen
ce in the ORs for premenopausal nonsmokers and smokers suggests an inc
reased risk for smokers. While there are limitations to this study, pa
rticularly related to the small number of subjects with the DC and CC
genotypes, the study suggests that some women may be susceptible to to
bacco smoke because of a CYP2E1 polymorphism. However, these results a
re preliminary and must be replicated. (C) 1996 Wiley-Liss, Inc.