ABSTRACT -As part of a case-control study in North Carolina involv- groups of women with differing risks (3): nonusers of to-
ing 206 women with oral and pharyngeal cancers and 352 controls, bacco, snuff users, and smokers who did not use snuff (few
questions were asked concerning the patterns of mouthwash use. No women both smoked and dipped snuff). Summary RR
significant overall increase in risk was found among users; the relative adjusting for tobacco habits and other indices and corre-
risk, adjusted for snuff dipping and smoking habits, was 1.15 [lower, sponding 95% CI were calculated according to the method
upper limits of the 95% confidence interval (95% CI)=0.8, 1.7]. The described by Gart (5) and Thomas (6).
relative risk associated with mouthwash use was increased to 1.94
(95% CI=0.8, 4.7), however, among women abstaining from tobacco. RESULTS
Although consistent dose-response relationships were not observed
for this subgroup, these findings and other reports of an increased risk
A total of 558 completed supplemental questionnaires
among persons ordinarily at low risk of this disease raise the possibility
were obtained, representing 89% of the cases and 86% of the
that mouthwash may contribute to oral and pharyngeal cancers.-JNCI
controls from whom a telephone interview was sought and
TABLE 1.-RR of oral and pharyngeal cancers associated with use TABLE 2.-RR of oral and pharyngeal cancers by years of
of mouthwash according to tobacco use and type ofrespondent mouthwash use according to type ofrespondent
Ever use Years Self interviews" Next-of-kin interviews"
mouthwash used
Tobacco Respondent Case Control RR b
on regular mouth-
Case Control RRc Case Control RRc
basis" wash
None Next of kin No 14 51 1.3 Nonuser 26 69 1.0 73 117 1.0
Yes 9 26 0-4 0 4 0.0 7 2 5.5
DK 3 7 5-9 0 4 0.0 4 4 1.2
10-24 4 12 1.0 13 10 2.0
Self No 1 33 8.3 ~25 13 48 0.7 18 30 0.9
Yes 7 28
DK 0 0 Excluded are 5 cases and 9 controls with unknown duration of use.
U
Self No 10 12 0.4
Yes 6 18 mouthwash among users of tobacco [the overall RR adjusted
TABLE 3.-RR of oral and pharyngeal cancers associated with years of mouthwash use, daily frequency of use, time usually held in the
mouth, and concentration, for selected tobacco and respondent categories
Tobacco abstainers
Snuff users: Next-of-kin interviews
Index Next-of-kin interviews Self interviews
who did not (37%). Controlling for denture use, however, It is not clear why an association with mouthwash use was
did not markedly change the RR in the various respondent detected only among women unexposed to tobacco and
and tobacco groups. In the group with the highest observed alcohol, the major risk factors for oral and pharyngeal
risk, directly interviewed women who used mouthwash but cancers (7, 8). Although chance is an obvious explanation,
not tobacco, only 1 of the 7 cancer patients had worn particularly in view of the small number of subjects in-
dentures. Mouthwash use also varied by education; the volved, the finding is consistent with two other studies
percentages of regular users were 45, 42, and 39% for control conducted in different areas of the United States (1, 2).
women with education levels of grade school, high school, Persons who abstain from tobacco and alcohol may share
and beyond high school, respectively. There was some var- with mouthwash users certain characteristics that increase
iation by age, with the highest percentage of users (49%) their risk of oral cancer, but few factors other than tobacco
among those of ages 60-69 years. Adjusting for educational and alcohol are known to be related to this cancer. Adjusting
or age differences also did not materially reduce the RR in for the confounding influences of denture wearing (9) and
the 3 groups with elevated risks. socioeconomic status, as measured by education level, al-
tered the findings only minimally. If mouthwash is indeed
DISCUSSION a risk factor for oral cancer, its effect may be comparatively
small, obscured by the more potent risk factors of tobacco
Two epidemiologic studies of oral and pharyngeal cancers
and alcohol and thus more easily detected in an otherwise
have reported that regular use of mouthwash may be an