Cancer Causesand Control 1,
235 - 241
Dentition, oral hygiene, and risk of oral
cancer: a case-control study in Beijing,
People's Republic of China
T o n g z h a n g Z h e n g , Peter Boyle, H u a n f a n g H u , J u n D u a n ,
Peijue J i a n g , D a q u a n Ma, L i a n g p e n g Shui, Shiru N i u ,
Crispian Scully, a n d Brian M a c M a h o n
(Received 24 July 1990; accepted in revised form 28 August 1990)
A case-control study of oral cancer was conducted in Beijing, People's Republic of China. The study was hospitalbased and controls were hospital in-patients matched to the cases by age and gender. A total of 404 case/control
pairs were interviewed. This paper provides data regarding oral conditions as risk factors for oral cancer, with every
patient having an intact mouth examined (pre-operation among cases) using a standard examination completed by
trained oral physicians. After adjustment for tobacco smoking and alcohol consumption, poor dentition--as reflected
by missing teeth--emerged as a strong risk factor for oral cancer: the odds ratio (OR) for those who had lost 15 - 32
teeth compared to those who had lost none was 5.3 for men and 7.3 for women and the trend was significant (P
< 0.01) in both genders. Those who reported that they did not brush their teeth also had an elevated risk (OR
= 6.9 for men, 2.5 for women). Compared to those who had no oral mucosal lesions on examination (OR -- 1.0),
persons with leukoplakia and lichen planus also showed an elevated risk of oral cancer among men and women.
Denture wearing per se did not increase oral cancer risk (OR = 1.0 for men, 1.3 for women) although wearing
metal dentures augmented risk (OR = 5.5 for men). These findings indicate that oral hygiene and several oral conditions are risk factors for oral cancer, independently of the known risks associated with smoking and drinking.
Key words: Alcohol, case-control study, dentition, leukoplakia, lichen planus, oral cancer, oral hygiene.
Introduction
Even though oral hygiene, inadequate dentition reflected
by loss of teeth, jagged teeth, septic and decayed teeth,
and oral mucosa lesions such as leukoplakia have long
been suspected to be risk factors for oral cancer, 1 10the
role of such variables has been overshadowed for a long
time by the results of studies relating tobacco smoking,
alcohol drinking and, in Eastern countries, betel chewing
to oral c a n c e r risk. 11'12 The independent and joint
-
effects of these factors with tobacco smoking and alcohol
drinking have not been studied extensively. Therefore,
a hospital-based case-control study was conducted in
Beijing to evaluate the possible role of oral hygiene and
dental conditions, as well as tobacco smoking, alcohol
drinking, diet, and nutrition in the genesis of oral cancer.
Here we report the data relating to oral hygiene and
dental conditions.
From the Department ofEpidemiology, National Institute ofEnvironmental Health and Engineering, Chinese Academy of Preventive Medicine,
Beijing, China (ZT; HH; NS); Unit of Analytical Epidemiology, International Agency for Research on Cancer, Lyon, France (PB; ZT); Belying
Union Hospital (DJ); Cancer Institute, Chinese Academy of Medical Science (]P); Beijing Medical University Stomatological Hospital (MD);
Bei]ing Municipal Stomatological Hospital (SL); UniversityDepartment of Oral Medicine and Oral Surgery, Br~tol Dental Hospital and School,
UK (CS); Department o f Epidemio/ogy, HarvardSchool of Public Health (BM; ZT). Address correspondenceto Dr Zheng at the CancerPrevention
Research Unit, Yale University, School of Medicine, 26 High Street, New Haven, CT 06510, USA. Dr Zheng was partly supported by a grant
from the DuPont Company.
© 1990 Rapid Communications of Oxford Ltd
235
T. Z/aeng et al.
Materials a n d m e t h o d s
A description of the study design has been given
previously. 13 Briefly, cases consisted of all the male and
female incident oral-cancer patients admitted to any one
of the seven hospitals that treat oral cancer in Beijing.
A total of 404 patients with histologically confirmed oral
cancer (ICD-9 TM codes 141 [cancer of the tongue] and
143 - 145 [other parts of the mouth, excluding salivary
gland; pharynx including nasopharynx]) were interviewed
between 1 May 1989 and 24 December 1990. The cases
ranged in age from 18 to 80 years at the time of
diagnoses. One control was selected for each case matched
on hospital, gender, and age within five years. Controls
were randomly chosen from patients whose current
hospitalizations were for minor conditions diagnosed
within one year and believed not to be associated with
the exposures of interest. The diagnoses in the controls
included minor surgery (e.g. hemorrhoids, strangulated
hernia, etc.), ophthalmic and ear conditions, low back
pain, urinary tract infection, and some other conditions.
There were no refusals to participate in either the cases
or their 404 first-selected controls.
A standard structured questionnaire was used to obtain
information on prior use of tobacco and alcohol, and
occupational and environmental exposures, as well as
other relevant factors. Interviews were conducted in
person by two trained, retired nurses who had no
knowledge of the study hypotheses. At each participating
hospital, there was one dentist who was responsible for
the histologic confirmation of all the cases, oral examination for all the cases and controls of that hospital, and
completion of a brief questionnaire, as an addendum to
the main questionnaire, including detailed questions on
dentition, oral hygiene, and the presence of oral mucosa
lesions. Questions included: how often the patients
visited a doctor because of toothache, oral ulceration, or
for a routine dental check-up; the frequency of tooth
brushing or denture cleaning; the number of years
elapsed between loss of teeth and having dentures fitted,
and the number of years of denture-wearing. Oral
examination included: recording of the number of
missing teeth, jagged teeth, filled teeth, decayed and
septic teeth; the presence of gingivitis or periodontal
disease and other oral mucosa lesions including
leukoplakia, erythroplasia and lichen planus. All the
interviews and oral examinations were conducted before
surgery. When a questionnaire was completed and sent
back to the research center by an interviewer, a quality
controller (who was also a dentist) would check the
accuracy of the coding and completeness of questionnaires. The correction of errors and collection of missing
information were undertaken by going back to the
hospital and re-interviewing the case or the control. As
236
with the interviewers, the doctors and quality controller
were not informed about the particular hypotheses being
evaluated.
The exposure odds ratio (OR) was used to estimate the
relative risk of oral cancer associated with each risk factor.
The crude OR (ORc) and the corresponding confidence
interval (CI) were calculated using the method of Mantel
and Haensze115 for each level of the exposure of interest.
In the text, numbers in parentheses following an OR are
95 percent CI of the OR.
The adjusted OR (ORa) was obtained from conditional
logistic regression analysis containing both the exposure
variable and potential confounding factors. 16 The
following potential confounding factors were included
in all the models: (i) alcohol drinking (using total lifetime
kgs of spirit-equivalent consumption in five categories
for men and three categories for women); (ii) tobacco
smoking (using pack-years of cigarette-equivalent
smoking in five categories for men and three categories
for women); (iii) years of education used as an indicator
of socioeconomic status in four categories for all analyses.
Gender and age were matched variables. When the
condition of the dentures worn and the risk of oral cancer
were assessed, the length of denture wearing in four
categories was also included in the model.
Formal statistical assessment of effect modification
between dentition, tobacco smoking, and alcohol
drinking was conducted by using conditional logistic
regression analyses controlling for education, gender, and
age. Tests for trend were performed by treating each
variable in the model as a continuous variable and the
ratio of the estimated coefficient to its standard error
obtained from the model was used to make statistical
inferfence based upon a two-sided alternative hypothesis
at the five percent level of significance.
Results
The crude and adjusted ORs for denture wearing and
the risk of oral cancer for men are given in Table 1. As
shown in this table, no significant association was
observed between oral cancer risk and the use of dentures,
either partial or complete. No association with oral cancer
risk was found either with years of denture wearing or
years elapsed between teeth extraction and having a
denture fitted. Regarding the type of material used in
the manufacture of the denture, more cases than controls
wore dentures made of metal (ORa of 5.5), but the
number of subjects who had this type of denture was
small and the excess not significant (CI 0.6 - 48.2). Table
2 presents the data for females. The same trends that were
seen among males are evident. No statistically significant association was found between wearing dentures and
the risk of oral cancer.
Oral cancer in China
Table 3 describes the association between the condition of the dentures and the risk of oral cancer. Even
though having a denture did not show any association
with oral cancer, the condition of the denture gave some
indication
women,
denture
of being
more
repaired
approximately
a risk f a c t o r . F o r b o t h
cases than
or
controls
rebased,
needed
with
two for having a denture
men
and
to have a
an
OR
of
which needed
T a b l e 1. D e n t u r e s a n d t h e risk o f oral cancer in m a l e s
Number
Risk factor
Cases
Controls
ORe a
ORa b
95% CI
186
62
49
13
184
64
52
12
i .0
1.0
0.9
1.1
1.0
1.0
1.0
0.9
Referent
0 . 6 - 1.6
0 . 6 - 1.8
0.3 - 2.3
37
11
14
38
13
13
1.0
0.8
1.1
1.0
0.8
1.2
0 . 6 - 1.8
0.3-2.1
0.5 - 2.9
4 - 10
> 10
15
27
20
18
25
21
0.8
1.1
0.9
1.0
1.0
0.9
0 . 5 - 2.0
0,5 - 2.2
0.4 - 2.1
Materials
Vulcanite
Plastic
Metal
0
52
10
0
62
2
0.8
5.0
0.9
5.5
0 . 6 - 1.5
0.6 - 48.2
Having denture
No
Yes
partial
complete
Interval c
1
2 -4
>4
Years d
1 - 3
aCrude OR.
bAdjusted OR. Adjusted by conditional logistic regression for tobacco smoking, alcohol drinking, years of education, gender, and age.
CNumber of years between loss of teeth and fitting of denture.
dNumber of years denture worn.
T a b l e 2. D e n t u r e s a n d t h e risk o f oral cancer in f e m a l e s
Number
Risk factor
Cases
Controls
ORe a
ORa b
95% CI
113
43
33
10
115
41
34
7
1.0
1.1
1.0
1.5
1.0
1.3
1.3
1.2
Referent
0.7 - 2.3
0.7 - 2.4
0.4 - 3.7
Interval c
1
2- 4
>4
27
7
9
22
11
8
1.3
0.7
1.1
1.5
0.8
1.5
0 . 7 - 3.2
0 . 2 - 2.4
0.4 - 4 . 8
Years d
1- 3
4 - 10
> 10
11
20
12
9
15
17
1.2
1.4
0.7
1.6
1.1
1.0
0 . 7 - 3.9
0 . 4 - 3.0
0,4 - 2.6
Materials
Vulcanite
Plastic
Metal
1
35
7
0
41
0
ND
0.9
ND
1.0
0 . 5 - 1.8
-
Having denture
No
Yes
partial
complete
acl!ade OR.
bAdjusted OR. Adjusted by conditional logistic regression for tobacco smoking, alcohol drinking, years of education, gender, and age.
CNumber of years between loss of teeth and fitting of denture.
dNumber of years denture worn.
ND, not determinable.
237
Ztseng et
T.
al.
T a b l e 3. D e n t u r e c o n d i t i o n a n d risk o f oral cancer
Number
Risk factor
Cases
Controls
ORc a
ORa b
95 % CI
MEN
Need repair
No
Yes
34
28
47
17
1.0
2.4
1,0
2.0
Referent
0,8 - 4.9
Need rebasing
No
Yes
31
31
41
23
1,0
1.8
1.0
1.8
Referent
0.8 - 4.3
WOMEN
Need repair
No
Yes
28
15
34
7
1.0
2.2
1.0
2.5
Referent
0,7 - 9.1
Need rebasing
No
Yes
26
17
33
8
1.0
2.2
1.0
3.2
Referent
0 . 9 - 11.4
~Crude OR.
bOR adjusted by conditionallogistic regression fortobacco smoking, alcohol drinking, education, number ofyears wearing denture, gender,
and age.
T a b l e 4. Loss o f t e e t h a n d t h e risk o f oral cancer in m a l e s
Number
Risk factor
Cases
Controls
ORc"
ORa b
95% CI
No lost teeth
46
99
1.0
1.0
Referent
Lost with
replacement
62
64
2.1
2.4
1.3 - 4.5
Lost without
replacement
140
85
3.5
3.7
2.2 - 6.4
38
59
55
50
60
36
24
29
1.4
1.3
0.6-
3.5
4.9
3.7
4.9
5.9
5.3
2.5
2 . 4 - 10,1
2 . 8 - 12.2
2 . 3 - 11.9
Number of lost
teeth with
replacement
1-2
3 -6
7 - 14
1 5 - 32
4
10
17
31
20
15
6
23
0.4
1.4
6.1
2.9
0,6
1.6
7.1
3.2
0.2-2.4
0.6-4.6
2.2-23.0
1 . 4 - 7.1
Number of lost
teeth without
replacement
1- 2
3-6
7 - 14
15 - 32
34
49
38
19
40
21
18
6
1.8
5.0
4.5
6.8
1.5
6.9
6.0
10.2
Number of
lost teeth
1- 2
3-6
7 - 14
15 - 32
~Crude OR.
boR adjusted by conditional logistic regression for tobacco smoking, alcohol drinking, years of education, gender, and age.
238
0.73.1 2.62.6 -
3.0
15.5
13.6
39.7
Oral cancer in China
to be repaired and/or rebased after adjustment for years
of denture wearing and other major confounding factors.
However, although all four ORs were high, each CI
included the value 1.0.
Table 4 presents the data relating loss of teeth to risk
of oral cancer for males. Having lost teeth, regardless of
replacement or not, was associated with increased risk for
oral cancer, with an OR of 2.4 (1.3 - 4.5) for the former
and 3.7 (2.2 - 6.4) for the latter. For the number of lost
teeth, an increased risk with increasing number of lost
teeth was noted, which is even more pronounced for
having lost teeth without replacement; tests for trends
are significant at P < 0.01. Table 5 presents the same
data for females; a similar trend but a much stronger
effect of loss of teeth on the risk of oral cancer was noted.
Oral hygiene, as reflected by whether teeth were
brushed, is strongly associated with the risk of oral cancer
(Table 6). An adjusted OR of 6.9 (2.5 - 19.4) for men
and 2.5 (0.9 - 7.5) for women were found for those not
brushing their teeth compared with those brushing their
teeth once or more per day.
The results in Table 7 suggest a strong association
between oral leukoplakia, lichen planus, and risk of oral
cancer. Among males, more than seven percent of the
cases were found to have leukoplakia compared with less
than 0.5 percent of the controls, and 10 percent of the
cases were found to have lichen ptanus compared with
less than one percent of the controls. For females, even
though six percent of the cases were found to have oral
leukoplakia and nine percent of the cases were found to
have lichen planus, no controls were found to have either
of the diseases. On the other hand, the great majority
of cases of oral cancer occurred in the absence of any of
these lesions.
Table 8 presents the associations of oral cancer with
inadequate dentition, tobacco smoking, and alcohol
drinking considered singly or in combination. In these
analyses, inadequate dentition was defined according to
whether a person had lost teeth, regardless of replacement. Smoking and drinking were also dichotomized
into 'yes' or 'no.' Therefore, the referent group for these
analyses was composed of persons who neither smoked
tobacco, nor drank alcohol, nor had lost any teeth. As
Table 8 shows, while smoking and drinking each carried
a risk of oral cancer of approximately two-fold, inadequate dentition alone carried almost four times the risk
of that of adequate dentition, with an OR of 3.9
(2.1 - 7.4). When the joint effects of all three risk factors
Table 5. Loss of teeth and the risk of oral cancer in females
Number
Risk factor
Cases
Controls
ORc"
ORa b
95 % CI
No lost teeth
23
67
1.0
1.0
Referent
Lost with
replacement
43
41
3.1
5.6
2.2 - 14.5
90
48
5.5
8.3
3.5- 19.6
2
6
14
32
29
36
35
33
22
23
20
24
3.8
4.6
5.1
4.0
5.2
8.8
10.4
7.3
1.93.2 3.5 2.5-
14.1
24.6
30.9
21.6
Number of lost
teeth with
replacement
1- 2
3- 6
7 - 14
15 - 32
4
8
13
18
3
8
13
17
3.9
2.9
2.9
3.1
5.0
6.6
7.7
4.7
0.61.4 1.91.5-
38.2
31.2
30.8
15.0
Number of lost
teeth without
replacement
1- 2
3- 6
7 - 14
15 - 32
25
28
22
15
19
15
7
7
3.8
5.4
9.2
6.2
4.8
8.9
13.1
16.4
Lost without
replacement
Number of lost
teeth
13715 -
1.7 - 13.9
3 . 1 - 25.3
3.6-48.0
3.3 - 80.7
~Crude OR.
BOR adjusted by conditional logistic regression for tobacco smoking, alcohol drinking, years of education, gender, and age.
239
T. Zbeng et al.
were examined, it was f o u n d that those who h a d all three
( i n a d e q u a t e d e n t i t i o n , tobacco s m o k i n g , a n d alcohol
d r i n k i n g ) carried a risk m o r e t h a n 15 times greater t h a n
those who h a d n o n e o f these risk factors.
T h e possible association b e t w e e n oral cancer risk a n d
the history o f oral ulceration, X-ray or radiation exposure
to the head or oral cavity, b r o k e n teeth, decayed teeth,
septic teeth, filled teeth, a n d caries was also e x a m i n e d .
N o n e of these factors presented itself as a noteworthy risk
factor.
Discussion
The findings o f this study are consistent with several other
recent e p i d e m i o l o g i c studies 17- 19 which suggest that
there is n o association b e t w e e n oral cancer risk a n d the
use o f d e n t u r e s , that there is an increased risk associated
with a n increased n u m b e r o f m i s s i n g teeth, a n d a
decreased risk associated with t o o t h b r u s h i n g . W e have
shown t h a t these associations are i n d e p e n d e n t o f use o f
tobacco a n d alcohol.
Table 6. Tooth-brushing, and risk of oral cancer
Number
Cases
Controls
ORca
ORab
95% CI
1.0
5.1
1.0
6.9
Referent
2.5 - 19.4
84
1.0
155
1.6
9
7.0
Linear trend test P = 0.004
1.0
1.2
7.8
Referent
0.7- 1.9
2.6-23.3
MEN
Tooth-brushing
Yes
No
Brushing times
per day
> 1
1
0
208
40
239
9
53
155
40
WOMEN
Tooth-brushing
Yes
No
Brushing times
per day
> 1
1
0
134
22
53
81
22
149
1.0
1.0
7
3.5
2.5
Referent
0.9-7.5
79
1.0
70
1.7
7
4.7
Linear trend test P = 0.11
1.0
1.1
2.7
Referent
0.6 - 2.0
0.9-8.3
aCrude OR.
bAdjusted OR. Adjusted by conditional logistic regression for tobacco smoking, alcohol drinking, years of education, gender, and age.
Table 7. Oral mucosal lesions and risk of oral cancer
Number
Cases
Controls
201
19
2
26
245
1
0
2
ORca
ORab
95% CI
1.0
23.2
ND
15.9
1.0
17.6
31.2
Referent
2.2 - 140.5
1.0
ND
ND
ND
1.0
ND
ND
ND
MEN
No lesions
Leukoplakia
Erythroplasia
Lichen planus
4.0 - 247.2
WOMEN
No lesions
Leukoplakia
Erythroplasia
Lichen planus
128
9
4
15
156
0
0
0
aCrude OR.
bAdjusted OR. Adjusted by conditional logistic regression for tobacco smoking, alcohol drinking, years of education, gender, and age.
ND, not determinable.
240
Oral cancer in China
Table 8. Inadequate dentition, smoking, drinking, and the risk of oral cancer
Inadequate
dentition
Alcohol
drinking
-
Tobacco
smoking
-
Number of
cases and
controls
108
28
46
225
53
-
+
+
+
+
-
+
46
+
+
+
119
183
ORaa
1.0
2.0
1.8
3.9
4.1
7.3
9.1
15.5
95% CI
Referent
0.7- 5.9
0.7-4.7
2.1 - 7.4
1.6- 10.4
2.9 - 18.1
4.4 - 19.0
7.2 - 33.7
aOR adjusted by conditional logistic regression for years of education, age, and gender.
( + ) risk factor present.
( - ) risk factor absent.
The relative risks in our study suggest that the interaction between these three factors is much more than
additive and is approximately multiplicative. All three
of these factors lead to chronic trauma to and irritation
of the oral mucosa and it is reasonable to suppose that
this may be the underlying mechanism of all three
associations. On the other hand, the possibility cannot
be ruled out there is some other unidentified factor(s)
whose action is facilitated by these variables. For the
present, however, programs for the prevention of oral
cancer should focus on all three of these factors--the
prevention of tooth loss, heavy alcohol consumption, and
tobacco use.
References
1. Martin H, Pflueger OH. Cancer of the cheek (buccal
mucosa). Arch Surg 1935; 30:731 - 4 7 .
2. Wynder EL, Bross IJ, Feldman R. A study of the etiology
of factors in cancer of the mouth. Cancer 1957; 6 :
1300 - 23.
3. Graham S, Dayal H, Rohrer T, et al. Dentition, diet,
tobacco and alcohol in the epidemiology of oral cancer.
JNCI 1977; 59:1611 - 8.
4. Silverman S, Gorsky M, Lazada F. Oral leukoplakia and
malignant transformation: a follow-up study of 257
patients. Cancer 1984; 53: 563- 8.
5. Browne RM, Camsey MC, Waterhouse JAH. Etiological
factors in oral squamous cell carcinoma. Community Dent
Oral Epidemiol 1977; 5:301 - 6.
6. Gupta PC, Mehta FS, Daftary DK, et al. Incidence rates
of oral cancer and natural history of oral precancerous lesion
in a 10-year follow-up study of India villagers. Community
Dent Oral Epidemiol 1980; 8:287 - 333.
7. Malaowalla AM, Silverman S, Mani NJ, et al. Oral cancer
in 57518 industrial workers of Gujarat, India. A prevalence
and follow-up study. Cancer 1976; 37: 1882- 6.
8. Silverman S, Bhargava R, Mani NJ, et al. Malignant
transformation and natural history of oral leukoplakia in
57518 industrial workers in Gujarat, India. Cancer 1976;
38: 1790- 5.
9. Pindborg JJ, Joist O, Renstrup G, et al. Studies in oral
leukoplakia: A preliminary report on the period prevalence
of malignant transformation in leukoplakia based on a
follow-up study of 248 patients. J A m Dent Assoc 1968;
78: 767- 71.
10. Silverman S, Rosen RD. Observations on the clinical
characteristics and natural history of oral leukoplakia. J A m
Dent Assoc 1968; 76: 772- 6.
11. Boyle P, Zheng T, Macfarlane GJ, et al. Etiology and
epidemiology of head and neck cancer. Current Opinion
in Oncology 1990; 2: 539- 45.
12. Boyle P, Macfarlane GJ, McGinn R, et al. Epidemiology
of head and neck cancer. In: De Vries N and Gluckman
J, eds. Second Primary Cancers o f Head and Neck.
Heidelberg: Springer-Verlag, 1990.
13. Zheng T, Boyle P, Hu H, etal. Cigarette smoking, alcohol
consumption, and risk of oral cancer: a case-control study
in Beijing, People's Republic of China. Cancer Causes and
Control 1990; 1 : 1 7 3 - 9 .
14. World Health Organization. International Classification
o f Diseases, 9th Revision. Geneva: WHO, 1974.
15. Mantel N, Haenszel W. Statistical aspects of the analysis
of data from retrospective studies of disease. JNCI 1959;
22: 719-48.
16. Breslow NE, Day NE. Statistical Methods in Cancer
Research. Vol L The Analysis of Case-controI Studies. Lyon:
International Agency for Research on Cancer, 1980; IARC
Sci Pub 32: 2 4 8 - 79.
17. Marshall J, Graham S, Haughey BP. Smoking, alcohol,
dentition and diet in the epidemiology of oral cancer.
Submitted for publication.
18. Kabat GC, Hebert JR, Wynder EL. Risk factors for oral
cancer in women. Cancer Res 1989; 49: 2803- 6.
19. Franco EL, Kowalski LP, Oliveira BV, et al. Risk factors
for oral cancer in Brazil: a case-control study, lntJ Cancer
1989; 43." 992 - 1000.
241