American Journal of Epidemiology

Copyright O 1997 by The Johns HopWns University School of Hygiene and Public Health
All rights reserved

Vol. 146, No. 2

Printed /h USA.

Diet and the Risk of Salivary Gland Cancer

Pamela L Horn-Ross,1 Mem'lee Morrow,1 and Britt-Marie Ljung2

antioxidants; ascorbic acid; cholesterol; diet; epidemiologic factors; salivary gland neoplasms

has shown vitamin A deficiency to be associated with

a significant increase in salivary gland tumors in rats
(14) and beta-carotene supplementation, but not retinol supplementation, to decrease the yield of dimethylbenzanthracene (DMBA)-induced salivary gland tumors (15). These findings suggest that investigation of
the relation between salivary gland cancer and diet,
particularly antioxidant vitamins, is warranted.

Cancer of the major salivary glands is relatively

rare, occurring slightly more often in men (1.2 per
100,000 men per year) than in women (0.7 per
100,000 women per year) (1). Little is known about
the etiology of salivary gland cancers. Only four casecontrol studies have been reported (2-5), with the
earliest study detecting no risk factors (2). The only
established risk factors for salivary gland tumors are
radiation exposure (3-10) and a history of any prior
cancer (3, 4).
The relation between diet and the development of
salivary gland cancer has not been examined previously in humans. However, the consumption of betacarotene and vitamin C has been found to be inversely
related to the development of epithelial tumors at a
number of sites, most notably those of the alimentary
and respiratory tracts (11-13). Experimental research

MATERIALS AND METHODS

A population-based case-control study was conducted in the greater San Francisco-Monterey Bay
area. All subjects were between the ages of 20 and 79
years, were residents of one of the nine counties comprising this region, spoke sufficient English to complete the interview, and had no prior history of salivary
gland cancer. Cases were identified through the
Greater Bay Area Cancer Registry, a population-based
cancer registry that is part of the Surveillance, Epidemiology, and End Results (SEER) program and the
statewide California Cancer Registry. Of 219 cases
diagnosed with a cancer of the major salivary glands
(ICD-O-1 codes 142.0-142.9) between July 1, 1989,
and June 30, 1993, 175 (75 percent) were interviewed.
Twenty-seven (14 percent) cases had died, seven (3

Received for publication February 19, 1997, and accepted for

publication March 26, 1997.
Abbreviation: ER, estrogen receptor.
1
Epidemiology Program, Northern California Cancer Center,
Union City, CA.
2
Department of Pathology, School of Medicine, University of
California, San Francisco, CA.
Reprint requests to Dr. Pamela L Horn-Ross, Epidemiology Program, Northern California Cancer Center, 32960 Alvarado-Niles
Road, Suite 600, Union City, CA 94587.

171

Downloaded from http://aje.oxfordjournals.org/ by guest on July 31, 2016

Cancer of the major salivary glands is relatively rare, and little is known about its etiology. The only
established risk factors are radiation exposure and a prior cancer. The role of diet in the development of
salivary gland tumors has not been addressed previously. The results from a population-based case-control
study conducted in the greater San Francisco-Monterey Bay area examining the association between dietary
intake and salivary gland cancer risk are presented. Of 199 cases diagnosed with salivary gland tumors
between 1989 and 1993, 150 (75%) were interviewed. Nine cases were subsequently excluded based on
review of pathology specimens. Of 271 controls identified through random-digit dialing and the Health Care
Finance Administration files, 191 (70%) were interviewed. Eight cases and seven controls who over- or
underreported dietary intake were excluded from analysis. Vitamin C intake of >200 mg/day compared with
<;100 mg/day was associated with a 60% decrease in salivary gland cancer risk (odds ratio (OR) = 0.40, 95%
confidence interval (Cl) 0.22-0.70). Inverse associations observed for carotene, vitamin E, and fiber from fruits
and vegetables were diminished when adjusted for vitamin C intake. Fiber from bean sources was associated
with a 51 % decrease in risk after adjusting for vitamin C intake (OR = 0.49, 95% Cl 0.26-0.92 for >1.4 g/day
compared with <0.4 g/day). Cholesterol intake was associated with elevated risk (OR = 1.67, 95% Cl 1.2-2.4
for a 10% increase in calories from cholesterol). These findings suggest that preventive strategies developed
for common chronic diseases, such as increased consumption of fruits and vegetables and limiting foods high
in cholesterol, also may be effective in preventing these rare tumors. Am J Epidemiol 1997;146:171-6.

172

Horn-Ross et ai.

exposure categories generally reflect tertiles among

controls.
RESULTS

The study population is described in table 1. The

mean age of subjects was 57.2 years; 43 percent were
women, and 78 percent were white. Among cases, 68
percent of tumors were in the parotid gland. Adenocarcinomas were the most common histologic type (40
percent) followed by mucoepidermoid tumors (29 percent).
Based on the limited experimental data available, of
primary interest in this study were the effects of antioxidant vitamins. In table 2, the associations between
several micronutrients and salivary gland cancer risk
are presented. No significant association between retinol intake and salivary gland cancer risk was observed. However, carotene intake was inversely associated with cancer risk at levels twice the US
TABLE 1. Distribution of demographic* and tumor
characteristics among persons with salivary gland cancer and
population controls, San Francisco-Monterey Bay area,
California, 1989-1993
Cases(n = 141)

Controls (n = 191)

No.

No.

Age (years)*
<4O
40-^*9
50-59
60-69
70-79

23
26
20
38
34

16
18
14
27
24

29
31
25
53
53

15
16
13
28
28

Sex
Male
Female

81
60

57
43

108
83

57
43

107
10
14
8
2

76
7
10
6
1

153
16
7
11
4

80
8
4
6
2

Location of tumor
Parotid gland
Submandibular gland
Sublingual gland
Multiple sites
Not specified

96
31
3
1
10

68
22
2
1
7

Histologic type of tumor

Adenocarctnoma
Mucoepidermoid
Acinarcell
Squamous cell
Malignant mixed
Other
Not reviewed

52
38
15
8
5
13
10

40
29
11
6

Race/ethnicity
White
Lata'na
African American
Asian
Partly Native American

4
10

* Mean aga for cases, 56.6 years (standard deviation 15.6) and
for controls, 57.6 years (standard deviation 15.4).

Am J Epidemiol

Vol. 146, No. 2, 1997

Downloaded from http://aje.oxfordjournals.org/ by guest on July 31, 2016

percent) refused to participate, and 15 (8 percent) were

not interviewed for other reasons. Permission to review pathologic material was obtained from 143 (95
percent) of the 150 persons interviewed. Specimens
were unavailable for three. Nine cases were subsequently excluded based on review of pathology specimens; six were considered to be metastatic lesions,
and three were benign. The results presented below are
based on 141 cases including the seven persons who
declined permission for specimen review and the three
for whom specimens were not available. The histologic data for these 10 cases were obtained from the
cancer registry.
Controls were identified through random-digit dialing with identification of older controls (aged 65-79
years) supplemented through the files of the Health
Care Finance Administration. A total of 271 eligible
controls were identified and frequency matched to
cases on 5-year age groups and sex; 191 (70 percent)
were interviewed. Thirty-eight (14 percent) refused
participation, and 42 (16 percent) were not interviewed for other reasons.
Telephone interviews were conducted using a standardized structured questionnaire. The interview included questions on a wide variety of topics. Dietary
intake during the year prior to diagnosis or interview
for cases and controls, respectively, was assessed using the 60-item National Cancer Institute Health Habits and History food frequency questionnaire (16) and
included an assessment of vitamin supplement use.
Dietary analyses exclude eight cases and seven controls who substantially under- or overreported their
intake, i.e., men aged <70 years reporting < 1,100 or
>4,200 calories per day, men aged ^70 years reporting <900 or >3,500 calories per day, women aged
<70 years reporting <700 or >3,000 calories per day,
and women aged ^ 7 0 years reporting <700 or
>2,5OO calories per day.
After initial examination of the data, odds ratios and
95 percent confidence intervals were estimated using
unconditional logistic regression analyses that controlled for age, sex, and caloric intake (17) (tables 2
and 3). In the analysis of macronutrients, the nutrient
density approach was used, i.e., the analysis included
both the macronutrient of interest divided by caloric
intake and caloric intake as separate variables in the
logistic models (18). When interpreting the effects of
multiple factors simultaneously and in determining the
best fitting, most informative models, Wald statistics,
the change in likelihood ratio statistics, and the magnitude of the change in the point estimates and confidence intervals were examined. Tests for trend reflect
change across categories of exposure. Cut-points for

Diet and the Risk of Salivary Gland Cancer

173

TABLE 2. Relation between mlcronutrient intake and salivary gland cancer risk among persons with
salivary gland cancer and population controls, San Frandsco-Monterey Bay area, California, 1989-1993
Model 1t

Cases
(n=133)

Controls
(n=184)

OR*

Retinol (pg/day)
375
376-620
>620

38
49
46

61
62
61

Carotene (pg/day)
2,100
2,102-3,900
>3,900

53
48
32

60
62
62

Mfcronutrtent

OR

95% Cl

1.0
1.3
1.2

0.72-2.3
0.68-2.3

1.0
1.3
1.2

0.73-2.4
0.64-2.2

1.0
0.82
0.54

0.47-1.4
0.29-0.99

1.0
0.86
0.64

0.49-1.5
0.33-1.2

Ptnd =

Vitamin C (mg/day)
100
101-200
>200

AWnd"0-22

0 0 5

68
38
27

63
59
62

1.0
0.58
0.40

0.34-1.0
0.22-0.70

49
50
34

63
60
61

1.0
1.0
0.69

0.59-1.8
0.38-1.2

ar.nd-0-22

1.0
1.1
1.2
Pu.nd"0-54

0.63-2.0
0.58-2.3

* OR, odds ratio; Cl, confidence interval.

t Adjusted for age, sax, and caloric intake.
% Adjusted for age, sex, caloric intake, and vitamin C intake (including supplements).
Including supplements.

recommended daily allowance, which is 1,500-2,000

/xg/day. Vitamin C intake (including supplements)
was also inversely related to risk at levels greater than
the recommended daily allowance (i.e., 60 mg/day or
^100 mg/day for smokers). When included in the
logistic model simultaneously, the effects for carotene
were diminished (table 2) whereas the effects for vitamin C were not substantially changed and remained
statistically significant. In addition, the change in the
likelihood ratio statistic associated with adding vitamin C intake to a model containing carotene intake
was statistically significant (x22 = 10.52, p 0.005);
however, the reverse was not observed (x12 = 2.38,
p = 0.30). We observed a nonsignificant inverse trend
for vitamin E intake (including supplements) that disappeared when we adjusted for vitamin C exposure.
Fruits and vegetables, the primary source of the
antioxidant vitamins of interest, are also a rich source
of fiber. Fiber from fruit and vegetable sources
showed a significant inverse relation with salivary
gland cancer risk (table 3). However, as with carotene,
this effect was diminished when we adjusted for vitamin C intake. Although fiber from grain sources was
not associated with risk, an inverse relation between
fiber from bean sources and salivary gland cancer risk
was observed. This effect was diminished only slightly
when we adjusted for vitamin C intake (table 2). The
Am J Epidemiol

Vol. 146, No. 2, 1997

change in the likelihood ratio statistic associated with

adding fiber from bean sources to a model containing
vitamin C intake approached but did not reach statistical significance (jf2 = 4.16, p = 0.12). Adjusting for
tobacco consumption did not alter the estimates of
effect for vitamin C consumption for men, women, or
all participants combined.
We also investigated the relation between macronutrients and salivary gland cancer risk. In crude dietary
analyses (i.e., unadjusted for energy intake), fat consumption was positively associated with risk in men
(odds ratio = 2.5, 95 percent confidence interval 1.05.8 for the highest compared with the lowest tertile).
This elevation in risk appeared to be due to the effects
of cholesterol consumption in amounts greater than the
recommended daily allowance, which is <300 mg/day
(odds ratio = 2.4, 95 percent confidence interval
0.94-6.1). Although no overall or saturated fat effect
was observed in women, cholesterol consumption was
associated with a significant increase in risk that was
similar in magnitude to that observed in men. When
we adjusted for caloric intake using a nutrient density
approach (18), a 10 percent increase in calories from
cholesterol was associated with a 67 percent increase
in cancer risk (95 percent confidence interval 1.2-2.4).
In table 4, the effects of vitamin C, fiber from beans,
and cholesterol are presented after adjustment for all

Downloaded from http://aje.oxfordjournals.org/ by guest on July 31, 2016

Vitamin E (mg/day)
7
7.1-14

Model 2%

95% Cl*

174

Hom-Ross et al.
TABLE 3. Relation between fiber intake and salivary gland cancer risk among persons with salivary
gland cancer and population controls, San Francisco-Monterey Bay area, California, 1989-1993
Source of ftoer
(0/day)
Fruits and vegetables
4.3
4.4-6.7
>6.7

(n= 133)

Controls
(n = 184)

OR

ModeMt
95% Cl*

58
40
35

62
61
61

1.0
0.67
0.56

Cases

42
47
44

59
69
56

95% Cl

0.39-1.2
0.31-1.0

1.0
0.76
0.75

0.43-1.3
0.39-1.4

0.56-1.7
0.61-2.0

1.0
1.1
1.4

0.61-1.9
0.74-2.6

).O5

Grains
2.5
2.6-4.3
>4.3

1.0
0.97
1.1

Ptrwd = 0- 2 4

/
Beans
0.4
0.5-1.4
>1.4

59
44
30

60
65
59

Model 2}
OR

1.0
0.71
0.49

1.0

0.66
0.46

0.39-1.1
0.25-0.84

0.42-1.2
0.26-0.92

TABLE 4. Relation between selected dietary components and salivary gland cancer risk among 133
persons with saltvary gland cancer and 184 population controls, San Francisco-Monterey Bay area,
California, 1989-1993
Dietary
component
vitamin C (mg/day)
100
101-200
>200

Model 2}

ModeMt
OR*

1.0
0.64
0.45
ftnwd "

Fiber from beans (g/day)

0.4
0.5-1.4
>1.4

1.0
0.73
0.53

Cholesterol (10% Increase)

1.6

95% Cl*

OR

95% Cl

0.34-1.2
0.21-0.96

1.0
0.74
0.49

0.40-1.3
0.25-0.94

002

Puma '

0.41-1.3
0.27-1.0

1.1-2.3

0 0 3

1.0
0.80
0.47

0.44-1.4
0.24-0.95

1.4

0.98-2.1

* OR, odds ratio; Cl, confidence interval.

t Adjusted for age, sax, caloric intake, retinol, carotene, vitamin E, fiber from fruits and vegetables, fiber from
grains, and the other factors in the table.
% Adjusted for age, sax, caloric Intake, therapeutic radiation to the head or neck, full mouth dental radiographic
examination, ultraviolet light treatment to the head or neck, occupational exposure to radiation or radioactive
materials, employment in the rubber industry, and occupational exposure to nickel compounds or alloys.
Including supplements.

the other dietary components of interest and nondietary factors associated with salivary gland cancer risk
in this study. Little evidence of confounding is observed, and the change in risk is generally less than 10
percent. Adjustment for race did not materially affect
the estimates of risk. Differences in risk factors by
histologic type were not generally apparent. The highest tertile of intake for fiber from grains was associated
with an elevation in risk of 2.3 (95 percent confidence
interval 0.88-5.9) for mucoepidermoid tumors. How-

ever, this analysis was based on 38 cases, and we

know of no biologically plausible reason for such an
association.
For men in this study, tobacco consumption was
associated with increased risk of developing an adenocarcinoma of the salivary gland (19). To the extent
possible, we examined the joint effects of tobacco and
vitamin C on the development of salivary gland cancer
in this subgroup. Risk associated with tobacco consumption was substantially greater among men conAm J Epidemiol

Vol. 146, No. 2, 1997

Downloaded from http://aje.oxfordjournals.org/ by guest on July 31, 2016

* OR, odds ratio; Cl, confidence interval.

t Adjusted for age, sex, and caloric intake.
% Adjusted for age, sex, caloric intake, and vitamin C Intake (including supplements).

Diet and the Risk of Salivary Gland Cancer

suming low levels of vitamin C than among those

consuming moderate/high levels (table 5). Note, however, that the numbers in this analysis are small, and
interpretation of this finding should be made with
caution.
DISCUSSION

associated with the regression of tobacco-induced carcinogenesis in in vitro studies (13). Several of the
observations from this study point to a possible role of
nitroso compounds in the development of salivary
gland cancer (19). These observations include the elevated risk associated with employment in the rubber
industry in which nitroso compounds are used; the
association with current cigarette use and the presence
of nitroso compounds in tobacco smoke; the protective
effects of vitamin C, which may be due to its ability to
block the formation of nitrosamines; and the possible
interaction between tobacco smoke and vitamin C
intake and the development of adenocarcinomas of the
salivary gland.
We observed a strong inverse relation between salivary gland cancer risk and fiber from bean sources.
These effects were diminished only slightly when we
adjusted for vitamin C intake. Fiber from grain
sources, however, was not associated with risk, nor
was fiber from fruits and vegetables once vitamin C
was taken into account. These findings lead to the
speculation that nonnutritive components of beans,
which are correlated with fiber content such as phytoestrogens (e.g., lignans) may be associated with diminished risk of salivary gland cancer. Phytoestrogens
are weak estrogens of plant derivation that may have
antiestrogenic effects through competitively binding to
estrogen receptors, thus diminishing the binding of
stronger endogenous estrogens (13, 21, 22). Because
of their weak estrogenic potential, phytoestrogens do
not elicit a major estrogenic response and thus have an
antiestrogenic effect that can inhibit the growth and
proliferation of hormone-dependent cancer cells (13,
21, 22). Dimery et al. (23) identified estrogen receptors (ERs) in both normal (79 percent) and cancerous
(86 percent) salivary gland tissue in men (78 percent
ER positive) and women (87 percent ER positive),
leading those authors to conclude that salivary gland
tumors may be hormonally dependent. Indeed, four of
the eight women with ER-positive tumors had salivary
gland estrogen receptor levels that were equivalent to

TABLE 5. Joint effects of cigarette smoking and vitamin C consumption among 29 men with
adenocardnoma of the salivary gland and 103 population controls, San Francisco-Monterey Bay area,
California, 1989-1993
Smoking
status

No
Vfes
No
\fes

VltaminC
Intake

Ever smokers

(matey)

95% Cl*

ORt

95% Cl

>100
>100
5100
5100

1.0
4.1
1.8
10.7

0.83-20.7
0.23-14.2
2.1-53.9

1.0
6.8
1.8
10.6

1.0-45.3
0.22-14.1
1.7-65.3

* OR, odds ratio; Cl, confidence interval,

t Adjusted for age, and caloric intake.

Am J Epidemiol

Current smokers

OR'.t

Vol. 146, No. 2, 1997

Downloaded from http://aje.oxfordjournals.org/ by guest on July 31, 2016

This is the only known study to evaluate the relation

between dietary intake and the risk of salivary gland
cancer in humans. Antioxidant vitamins, particularly
vitamin C, and beans were associated with lowered
risk and cholesterol, with increased risk. Consistent
with the limited experimental data (15), this study
found no significant association between retinol intake
and salivary gland cancer risk; but a significant inverse
relation was found between carotene intake and risk.
However, when adjusted for each other, the effects for
carotene were diminished whereas the effects for vitamin C were not substantially changed and remained
statistically significant. This pattern has been observed
in a number of studies assessing the relation between
vitamin C and carotenoids in the development of cancers of the oral cavity, larynx, esophagus, and lung
(11), suggesting that vitamin C may play a larger role
in cancer prevention than carotene for alimentary and
respiratory cancers. However, the nature of this anticarcinogenic effect remains to be determined.
In addition to its antioxidant properties, vitamin C
may have anticarcinogenic effects through its ability to
block the formation of nitrite, thus preventing the
formation of nitrosamines (13, 20). Nitrates found in
certain vegetables and drinking water are converted to
nitrite in the saliva and stomach. It remains to be
determined whether nitrosamines affect the risk of
salivary gland cancer; however, they have been associated with an increase in nasopharyngeal and esophageal tumors in rodents (13). In addition, nitrosamines
are major carcinogenic agents found in smoke from
cigarettes. Antioxidant vitamins have been found to be
protective against the development of a number of
tobacco-related cancers (20), and vitamin C has been

175

176

Horn-Ross et al.

those considered "hormonally dependent" in breast

tumors (i.e., 10 fmol/mg). ER levels in two of the four
men with ER-positive tumors were >7 fmol/mg and
thus approached those considered hormonally dependent.
The findings from the present study suggest that
preventive strategies developed for common chronic
diseases may be effective also in preventing salivary
gland cancer. Specifically, increasing consumption of
fruits and vegetables, particularly those high in vitamin C, and limiting foods high in cholesterol may be
effective in preventing such rare tumors.

7.
8.
9.
10.
11.
12.
13.
14.

This research was supported by grant R29-CA49499

from the National Cancer Institute.

15.
16.

REFERENCES
1. Ries LAG, Miller BA, Hankey BF, et al, eds. SEER cancer
statistics review, 1973-1991: tables and graphs. Bethesda,
MD: National Cancer Institute, 1994. (NIH publication no.
94-2789).
2. Keller AZ. Residence, age, race, and related factors in the
survival and associations with salivary tumors. Am J Epidemiol 1969;90:269-7.
3. Spitz MR, Tilley BC, Batsakis JG, et al. Risk factors for major
salivary gland carcinoma: a case-comparison study. Cancer
1984;54:1854-9.
4. Preston-Martin S, Thomas DC, White SC, et al. Prior exposure
to medical and dental x-rays related to tumors of the parotid
gland. J Natl Cancer Inst 1988;80:943-9.
5. Spitz MR, Fueger JJ, Goepfert H, et al. Salivary gland cancer
a case-control investigation of risk factors. Arch Otolaryngol
Head Neck Surg 1990;l 16:1163-6.
6. Belsky JL, Takeichi N, Yamamoto T, et al. Salivary gland

17.
18.
19.
20.
21.
22.
23.

Am J Epidemiol

Vol. 146, No. 2, 1997

Downloaded from http://aje.oxfordjournals.org/ by guest on July 31, 2016

ACKNOWLEDGMENTS

neoplasms following atomic radiation: additional cases and

reanalysis of combined data in a fixed population, 1957-1970.
Cancer 1975;35:555-9.
Maxon HR, Saenger EL, Buncher CR, et al. Radiationassociated carcinoma of the salivary glands: a controlled
study. Ann Otolaryngol 1981;90:107-8.
Modan B, Baidatz D, Mart H, et al. Radiation-induced head
and neck tumors. Lancet 1974; 1:277-9.
Ju DMC. Salivary gland tumors occurring after radiation of
the head and neck area. Am J Surg 1968;116:518-23.
Palmer JA, Mustard RA, Simpson WJ. Irradiation as an etiologic factor in tumors of the thyroid, parathyroid, and salivary
glands. Can J Surg 1980;23:39-42.
Block G. Vitamin C and cancer prevention: the epidemiologic
evidence. Am J Clin Nutr 1991;53:270S-82S.
Steinmetz KA, Potter JD. Vegetables, fruit, and cancer. I:
Epidemiology. Cancer Causes Control 1991 ;2:3257.
Steinmetz KA, Potter JD. Vegetables, fruit, and cancer. II:
Mechanisms. Cancer Causes Control 1991;2:427-42.
Rowe NH, Grammer FC, Watson FR, et al. A study of
environmental influence upon salivary gland neoplasia in rats.
Cancer 1970;26:436-44.
Alam BS, Alam SQ, Weir JC Jr, et al. Chemopreventive
effects of /3-carotene and 13-cis-retinoic acid on salivary gland
tumors. Nutr Cancer 1984;6:4-12.
Block G, Hartman AM, Naughton D. A reduced dietary
questionnaire: development and validation. Epidemiology
1990; 1:58-64.
Kelsey JL, Thompson WD, Evans AS. Methods in observational epidemiology. New York, NY: Oxford University
Press, 1986.
Willett W. Nutritional epidemiology. New York, NY: Oxford
University Press, 1990:245-71.
Horn-Ross PL, Ljung B-M, Morrow M. Environmental factors
and the risk of salivary gland cancer. Epidemiology (In press).
Block G. The data support a role for antioxidants in reducing
cancer risk. Nutr Rev 1992;50:207-13.
Rose DP. Dietary fiber, phytoestrogens, and breast cancer.
Nutrition 1992,8:47-51.
Adlercreutz H, Mousavi Y, Clark J, et al. Dietary phytoestrogens and cancer in vitro and in vivo studies. J Steroid Biochem Molec Biol 1992;41:331-7.
Dimery IW, Jones LA, Verjan RP, et al. Estrogen receptors in
normal salivary gland and salivary gland carcinoma. Arch
Otolaryngol Head Neck Surg 1987; 113:1082-5.