PRESERVED FOODS IN RELATION TO RISK OF NASOPHARYNGEAL

CARCINOMA IN SHANGHAI, CHINA

Jian-Min YUAN
1
*, Xue-Li WANG
2
, Yong-Bing XIANG
2
, Yu-Tang GAO
2
, Ronald K. ROSS
1
and Mimi C. YU
1
1
Department of Preventive Medicine, USC/Norris Comprehensive Cancer Center, University of Southern California School of Medicine,
Los Angeles, CA, USA
2
Department of Epidemiology, Shanghai Cancer Institute, Shanghai, Peoples Republic of China
A population-based case-control study was conducted in
Shanghai, China, to investigate the association between di-
etary factors and risk of nasopharyngeal carcinoma (NPC).
Thestudyincluded935NPC patientsaged15to74yearsand
1,032communitycontrols. Exposurestosaltedshandother
protein-containing preserved food were associated with in-
creased risk of NPC. Individuals who ate salted sh at least
once a week had an 80%increase in risk of NPC relative to
thosewho atesaltedshlessthanonceamonth(p 0.07).
Compared with those in the lowest quartile of protein-
containingpreservedfoods, subjectsinthehighest quartileof
intake experienced a statistically signicant 78%increase in
riskof NPC [oddsratio(OR) 1.78, 95%condenceinterval
(CI) 1.372.31], withadose-dependent relationship(p for
linear trendF0.001). A similar associationbetweenintakeof
preservedvegetablesandNPC riskwasobserved(OR1.39,
p for linear trend 0.003). In contrast, high intake of
oranges/tangerineswasassociatedwithastatisticallysigni-
cant reductioninriskofNPC (OR0.55, pfor linear trendF
0.001). Whenweexaminedthejoint effect of preservedfood
andoranges/tangerinesonriskofNPC,subjectsinthehighest
tertile of preserved food and the lowest tertile of orange/
tangerine intake had a 3-fold increase in risk (95%CI
2.084.91) compared with those in the lowest tertile of
preserved food and the highest tertile of orange/tangerine
intake. Int. J. Cancer 85:358363, 2000.

2000 Wiley-Liss, Inc.

Although rare in most other parts of the world, nasopharyngeal
carcinoma (NPC) is a common malignancy in southern China,
especially for the Cantonese who inhabit the central region of
Guangdong Province (National Cancer Control Office, 1979; Yu et
al., 1981). Ingestion of Chinese salted sh has been found to be an
important risk factor for NPC among these high-risk Cantonese
Chinese as well as among intermediate- to low-risk Chinese
populations of non-Cantonese origin (Chen et al., 1988; Ning et al.,
1990; Yu et al., 1986, 1989a; Y.M. Zheng et al., 1994; Yu and
Henderson, 1996). Age at rst exposure to salted sh has been
established to be an important co-determinant of risk in both high-
and low-risk Chinese populations; risk increases as age at rst
exposure decreases (Ning et al., 1990; Yu et al., 1986, 1989a). In
fact, salted sh is the most common weaning food in Cantonese,
who have the highest reported incidence of the disease in the world
(Yu et al., 1986, 1989a). Besides salted sh, exposures to other
preserved food products, such as salted shrimp paste, fermented
soybean paste and various types of salted vegetable, have been
shown to be related to NPC risk in the Chinese (Chen et al., 1988;
Ning et al., 1990; Yu et al., 1988, 1989a; Yu and Henderson, 1996).
Amajor focus of the present study was to investigate the possible
roles of Chinese salted sh and other preserved foods in the
development of NPC in Shanghai, China, an intermediate-risk
population for this malignancy (4.0 per 100,000 person-years in
men, 1.9 per 100,000 person-years in women) (Parkin et al., 1992).
The population-based case-control study included 935 incident
cases of NPC and 1,032 community controls. This report presents
our dietary ndings. A separate article describes the results of
non-dietary risk factors for NPC in Shanghai, China (Yuan et al.,
1999).
MATERIAL AND METHODS
Subjects
NPC patients were identied through the Shanghai Cancer
Registry, a population-based registry covering the approximately 7
million residents of the city of Shanghai, Peoples Republic of
China. The registry identied 1,110 patients aged 15 to 74 years
with histologically conrmed NPC between January 1987 and
September 1991. Among them, 87 patients died before we could
contact them and 8 refused to be interviewed. We were unable to
locate 80 patients. Thus, we interviewed a total of 935 (84%)
eligible patients.
Control subjects were randomly selected from the urban Shang-
hai population and chosen to match the frequency distribution by
sex and 5-year age groups of NPC patients as ascertained by the
Shanghai Cancer Registry during 19801984. Personal identica-
tion cards were used to select potential control subjects. These
cards, one per resident, were housed in 4,410 le cabinet drawers
(which are numbered from 1 to 4,410 at the Residents Registry,
Bureau of Public Security of Shanghai). We rst generated 1,070
random numbers between 1 and 4,410. The number 1,070 was our
anticipated number of incident NPC cases during the proposed 4
years and 9 months of the study period. Our goal was to recruit
approximately equal numbers of cases and controls. The 1,070
random numbers we generated became the designated drawer
numbers from which identication cards for the 1,070 eligible
control subjects were selected according to a standard algorithm.
Only 38 eligible control subjects refused to participate in the study,
resulting in a nal sample size of 1,032 control subjects.
Data collection
All interviews were conducted in person by 4 trained interview-
ers employing a structured questionnaire. The questionnaire re-
quested information up to 2 years prior to the diagnosis of cancer
for cases and 2 years prior to the date of interview for controls
(reference date). The questionnaire included demographic charac-
teristics, usual dietary habits during adulthood, use of tobacco and
alcohol, lifetime exposure to secondhand cigarette smoke, types of
oil and fuel used in cooking, lifetime occupational history, history
of chronic ear and nose conditions and family history of NPC.
In terms of usual adult dietary habit, subjects were asked to
estimate the intake frequency (times per day, per week, per month
or per year) of each of 38 specic food groups (see Appendix). For
seasonal foods which included most vegetables and fruits, subjects
were asked to estimate the intake frequencies while the food items
were in season.
All interviews were conducted between January 1988 and
December 1991. Eighty percent of cases were interviewed within 6
Grant sponsor: National Cancer Institute; Grant numbers: R01 CA40468;
R35 CA53890.
*Correspondence to: Department of Preventive Medicine, USC/Norris
Comprehensive Cancer Center, M/S #44, University of Southern Califor-
nia, 1441 Eastlake Avenue, Los Angeles, CA 90033, USA. Fax: 1323
8650136. E-mail: jyuan@hsc.usc.edu
Received 15 June 1999; Revised 12 August 1999
Int. J. Cancer: 85, 358363 (2000)

2000 Wiley-Liss, Inc.

Publication of the International Union Against Cancer
months following the diagnosis of NPC (mean 3.5 months).
Approximately an equal number of controls per year were enrolled
into the study during the 4-year recruitment period.
Data analysis
Case-control data were analyzed by standard methods (Breslow
and Day, 1980). Unconditional logistic regression models were
used to examine the association between exposure to dietary
factors and NPC risk. Exposuredisease associations were mea-
sured by odds ratios (ORs) and their corresponding 95% con-
dence intervals (CIs) and p values. Linear trend tests for exposure
disease associations were based on categorical variables. Exposure
indices for broader food groups were constructed by adding the
intake frequencies of individual food groups. For example, intake
frequencies (expressed in numbers of times per year) of salted eggs
and fermented black eggs were summed to generate an exposure
index for preserved eggs. Intake frequencies were grouped as less
than monthly (1/month), monthly (1/month to 1/week), weekly
(1/week to 1/day) or daily (1/day). For certain combined food
groups, individuals were categorized into tertiles or quartiles based
on the distribution of intake frequencies among control subjects.
Age (40, 4049, 5059, 60 years), gender (for both sexes
combined only) and level of education (primary school or less vs.
middle school or higher) were included as co-variates in all
analyses. These unadjusted analyses were repeated with the
inclusion of other risk factors for NPC identied in the present
study (Yuan et al., 1999): cigarette smoking (lifelong non-smoker,
20 cigarettes/day, 20 cigarettes/day), exposure to smoke from
heated rapeseed oil (yes vs. no) and burning coal (yes vs. no) during
cooking, occupational exposure to chemical fumes (yes vs. no) and
history of chronic ear and nose conditions (yes vs. no). Family
history of NPC (i.e., one or more rst-degree relatives with NPC), a
risk factor for NPC in the present study, was not adjusted for in our
dietary analyses since diet is part of shared environmental factors
within families and may be at least partially responsible for the
observed familial correlation. Analyses were conducted for men
and women separately and for both sexes combined. Cross-product
terms were included in logistic regression models to examine
genderexposure interaction effects on NPC risk. Results from the
unadjusted and adjusted analyses were similar; adjusted results are
presented. All p values are 2-sided.
RESULTS
Salted sh consumption was relatively rare in the study popula-
tion. In both sexes combined, only 20 NPC patients (2%) and 11
control subjects (1%) ate salted sh at least once a week. There was
a positive association between intake frequency and risk that was of
borderline statistical signicance (p for linear trend 0.07). Other
types of protein-containing preserved food also were positively
related to NPC risk. The association reached statistical signicance
for fermented beans/pastes (p for linear trend 0.001). Overall,
intake frequency of protein-containing preserved foods exhibited a
highly signicant dose-response relationship with NPC risk after
adjustment for other risk factors (p for linear trend 0.001) (Table I).
No difference in risk between the 2 sexes was detected (p 0.92).
Table II presents the association between intake frequencies of
various preserved vegetables and risk of NPC. High intake of
preserved stem vegetables (harm wou sun, cheung wou sun and
cheung wong qua) was associated with a statistically signicant
increase in risk of NPC. There was a signicant increase in risk
TABLE I INTAKE FREQUENCIES OF PROTEIN-CONTAINING PRESERVED FOODS IN RELATION TO RISK OF NPC IN SHANGHAI, CHINA
1
Total Men Women
Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI)
Salted sh
Less than monthly 802/913 1.0 578/620 1.0 224/293 1.0
Monthly 113/108 1.17 (0.881.57) 80/73 1.18 (0.831.67) 33/35 1.18 (0.702.00)
Weekly or more 20/11 1.82 (0.863.88) 10/6 1.45 (0.514.08) 10/5 2.29 (0.766.93)
p for linear trend 0.07 0.26 0.15
Salted seafood pastes
No 873/984 1.0 632/668 1.0 241/316 1.0
Yes 62/48 1.44 (0.972.15) 36/31 1.23 (0.742.05) 26/17 1.79 (0.933.45)
p 0.07 0.42 0.08
Preserved meats
Less than monthly 624/679 1.0 442/447 1.0 182/232 1.0
Monthly 257/319 0.87 (0.711.06) 191/224 0.87 (0.681.10) 66/95 0.87 (0.591.28)
Weekly or more 54/34 1.77 (1.122.79) 35/28 1.33 (0.782.24) 19/6 3.88 (1.4710.22)
p for linear trend 0.56 0.87 0.20
Preserved eggs
Less than monthly 249/336 1.0 179/231 1.0 70/105 1.0
Monthly 522/527 1.34 (1.091.66) 371/344 1.38 (1.071.78) 151/183 1.26 (0.861.86)
Weekly or more 164/169 1.17 (0.881.55) 118/124 1.07 (0.761.49) 46/45 1.51 (0.892.55)
p for linear trend 0.11 0.38 0.11
Fermented beans/pastes
Less than monthly 432/559 1.0 313/374 1.0 119/185 1.0
Monthly 257/275 1.25 (1.001.55) 177/193 1.12 (0.861.45) 80/82 1.60 (1.072.39)
Weekly or more 246/198 1.76 (1.392.22) 178/132 1.73 (1.312.29) 68/66 1.85 (1.202.85)
p for linear trend 0.001 0.001 0.002
All protein-containing preserved
foods (times/years)
1st quartile (27) 180/268 1.0 131/176 1.0 49/92 1.0
2nd quartile (2751) 228/252 1.34 (1.021.75) 163/174 1.26 (0.921.74) 65/78 1.51 (0.922.47)
3rd quartile (5296) 219/258 1.19 (0.911.56) 157/174 1.15 (0.831.58) 62/84 1.34 (0.812.20)
4th quartile (97) 308/254 1.78 (1.372.31) 217/175 1.63 (1.192.22) 91/79 2.17 (1.353.50)
p for linear trend 0.001 0.006 0.004
1
See Appendix for denitions of various food groups; intake frequency was categorized as follows: less than monthly, 1/month; monthly,
1/month to 1/week; weekly or more, 1/week.
2
Ca/Co, number of cases/number of controls.
3
OR, odds ratio, adjusted for age, gender (for
total only), level of education, cigarette smoking, exposure to smoke from heated rapeseed oil and burning coal during cooking, occupational
exposure to chemical fumes and history of chronic ear and nose condition (see text for more detailed description of confounding variables); CI,
condence interval.
359
DIETAND NPC
with increasing intake frequency in men and women separately and
both sexes combined. An elevated risk of NPC was observed for
intake of preserved leafy and root vegetables in women but not in
men. Overall, subjects in the highest quartile of preserved veg-
etable intake experienced a statistically signicant 43% increase in
risk of NPC relative to those in the lowest quartile. Risk of NPC
increased with increasing frequency of intake (p for linear trend
0.001). A slightly stronger association between intake frequency
and NPC risk was observed in women than in men, but the
difference was not statistically signicant (p 0.13).
We then summed the intake frequencies of all preserved foods to
create an overall preserved food exposure index. There was a
monotonic relationship between dose and NPC risk (Table II).
Subjects in the highest quartile of intake had a statistically
signicant 39% increase in NPC risk relative to those in the lowest
quartile of intake after adjustment for other potential risk factors for
NPC. A highly statistically signicant dose-response relationship
was present among women and both sexes combined; this relation-
ship was of borderline statistical signicance in men (p for linear
trend 0.08) (Table II).
Table III shows the association between intake of fresh veg-
etables and fruit and risk of NPC. Three specic types of vegetable,
green beans/peas, green/red peppers and tomatoes, demonstrated
statistically signicant inverse, dose-dependent relationships with
NPC risk in women (p values for linear trends: green beans/peas,
0.002; green/red peppers, 0.04; tomatoes, 0.02) but not in men
(corresponding p values were 0.38, 0.62 and 0.29, respectively). A
signicant inverse dose-response relationship between overall
intake of fresh vegetables and NPC risk was observed in women
(OR 0.63, p for linear trend 0.03) but not in men (OR 0.97,
p for linear trend 0.71) (Table III).
There was only one fresh fruit item, oranges/tangerines, in the
questionnaire. The risk of NPC was reduced at high intake of
oranges/tangerines. Subjects who ate this fruit at least once a week
while it was in season experienced a statistically signicant 45%
reduction in NPC risk relative to those who ate this fruit less than
once a month. Risk of NPC decreased with increasing frequency of
intake in men and women separately and in both sexes combined
(all p values for linear trend were 0.01) (Table III).
We examined the inter-related effects of oranges/tangerines,
green beans/peas, green/red peppers and tomatoes on NPC risk.
The effects of the 3 fresh vegetables were explained by that of
oranges/tangerines. The adjusted ORs were 0.82 for daily intake of
green beans/peas (p for linear trend 0.10), 0.67 for green/red
peppers (p for linear trend 0.43) and 0.92 for tomatoes (p for
linear trend 0.10) relative to less than weekly intake. In contrast,
after adjustment for the 3 fresh vegetables, the effect of oranges/
tangerines on risk of NPC remained statistically signicant (ad-
justed OR 0.61 for daily vs. less than weekly intake, p for linear
trend 0.0005).
There was little difference between cases and controls in intake
frequency of preserved fruits (Chan pai mui, gar ink gee, gone hong
zhao and bei guo) and fermented sauces used in avoring (see yao,
yu lo and ho yao). Similarly, NPC patients did not differ signi-
cantly from control subjects in consumption of fresh organ meats,
eggs, seafood and milk.
We examined the joint effect of intake of preserved food and of
oranges/tangerines on risk of NPC (Table IV). Risk of NPC
increased with increasing intake frequency of preserved food
within each intake level of oranges/tangerines. Similarly, an
inverse association between orange/tangerine intake and risk of
TABLE II INTAKE FREQUENCIES OF PRESERVED VEGETABLES IN RELATION TO RISK OF NPC IN SHANGHAI, CHINA
1
Total Men Women
Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI)
Preserved leafy vegetables
Less than monthly 367/404 1.0 271/269 1.0 96/135 1.0
Monthly 404/465 0.93 (0.761.13) 289/324 0.86 (0.681.10) 115/141 1.07 (0.731.56)
Weekly 147/149 0.94 (0.711.25) 101/97 0.89 (0.641.25) 46/52 1.03 (0.631.71)
Daily 17/14 1.37 (0.652.90) 7/9 0.75 (0.272.10) 10/5 2.90 (0.909.34)
p for linear trend 0.85 0.29 0.31
Preserved stem vegetables
Less than monthly 295/422 1.0 219/291 1.0 76/131 1.0
Monthly 222/271 1.15 (0.911.46) 159/184 1.12 (0.851.49) 63/87 1.26 (0.801.97)
Weekly 353/291 1.67 (1.332.08) 248/189 1.67 (1.282.17) 105/102 1.68 (1.112.52)
Daily 65/48 1.88 (1.242.84) 42/35 1.52 (0.922.49) 23/13 3.06 (1.436.55)
p for linear trend 0.001 0.001 0.001
Preserved root vegetables
Less than monthly 162/200 1.0 121/142 1.0 41/58 1.0
Monthly 272/353 0.94 (0.701.23) 206/234 1.01 (0.741.38) 66/119 0.77 (0.461.30)
Weekly 423/401 1.21 (0.941.57) 295/274 1.18 (0.871.59) 128/127 1.27 (0.782.08)
Daily 78/78 1.04 (0.701.53) 46/49 0.85 (0.521.39) 32/29 1.37 (0.702.67)
p for linear trend 0.14 0.61 0.06
All preserved vegetables (times/
year)
1st quartile (41) 194/258 1.0 153/172 1.0 41/86 1.0
2nd quartile (41102) 193/264 0.91 (0.691.19) 147/192 0.82 (0.601.12) 46/72 1.20 (0.702.07)
3rd quartile (103200) 237/253 1.18 (0.911.54) 163/172 1.05 (0.761.43) 74/81 1.60 (0.972.67)
4th quartile (201) 311/257 1.43 (1.111.86) 205/163 1.24 (0.911.69) 106/94 2.03 (1.253.28)
p for linear trend 0.001 0.07 0.002
All preserved foods (times/year)
1st quartile (87) 187/261 1.0 144/179 1.0 43/82 1.0
2nd quartile (87171) 200/255 1.03 (0.791.35) 152/176 0.99 (0.721.36) 48/79 1.16 (0.681.97)
3rd quartile (172334) 266/260 1.32 (1.021.72) 180/178 1.18 (0.861.61) 86/82 1.77 (1.082.90)
4th quartile (335) 282/256 1.39 (1.071.81) 192/166 1.27 (0.931.73) 90/90 1.75 (1.082.85)
p for linear trend 0.003 0.08 0.009
1
See Appendix for denitions of various food groups; intake frequency was categorized as follows: less than monthly, 1/month; monthly,
1/month to 1/week; weekly, 1/week to 1/day; daily, 1/day.
2
Ca/Co, number of cases/number of controls.
3
OR, odds ratio, adjusted for age,
gender (for total only), level of education, cigarette smoking, exposure to smoke from heated rapeseed oil and burning coal during cooking,
occupational exposure to chemical fumes and history of chronic ear and nose condition (see text for more detailed description of confounding
variables); CI, condence interval.
360
YUAN ET AL.
NPC was seen regardless of the intake level of preserved food. The
highest relative risk for NPC (OR 3.19, 95% CI 2.084.91) was
noted among subjects in the highest tertile of preserved food and
the lowest tertile of orange/tangerine intake (Table IV).
DISCUSSION
Chinese salted sh has been signicantly associated with NPC in
virtually every epidemiological study which has investigated this
possible exposuredisease association (Armstrong et al., 1983,
1998; Chen et al., 1988; Geser et al., 1978; Henderson et al., 1976;
Ning et al., 1990; Yu et al., 1986, 1988, 1989a; Y.M. Zheng et al.,
1994). Moreover, rats fed this human food develop cancer of the
nasal cavity, which otherwise rarely occurs in this species (Huang
et al., 1978; Yu et al., 1989b; X. Zheng et al., 1994). Low levels of
several nitrosamines have been detected in samples of Chinese
salted sh. Most of these volatile nitrosamines are capable of
inducing nasal cavity tumors in experimental animals (Yu and
Henderson, 1996). In addition, other bacterial mutagens and
genotoxic substances have been found in extracts of this food (Yu
and Henderson, 1996). Our data indicate that even though consump-
tion of salted sh is a risk factor for NPC in Shanghai, China, the
proportion of NPC cases attributable to salted sh in this popula-
tion is relatively small.
Besides salted sh, exposure to other preserved foods, whether
proteincontaining or of plant origin, was found to be associated
with an increased risk of NPC in Shanghai. This, too, is a fairly
consistent nding across populations. For example, we had previ-
ously noted independent effects of moldy bean curd, salted shrimp
paste, salted eggs and various preserved vegetables on NPC risk in
different Chinese populations (Yu et al., 1986, 1988, 1989a; Ning
TABLE III INTAKE FREQUENCIES OF FRESH VEGETABLES AND FRUIT IN RELATION TO RISK OF NPC IN SHANGHAI, CHINA
1
Total Men Women
Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI)
Bok choi
Less than weekly 10/18 0.56 (0.251.26) 6/11 0.52 (0.191.47) 4/7 0.60 (0.172.16)
Weekly 415/450 1.0 293/319 1.0 122/131 1.0
Daily 510/564 0.97 (0.811.17) 369/369 1.07 (0.861.33) 141/195 0.77 (0.551.08)
p for linear trend 0.85 0.32 0.24
Other dark green leafy vegetables
Less than weekly 123/167 1.0 79/118 1.0 44/49 1.0
Weekly 775/832 1.12 (0.861.46) 561/560 1.35 (0.981.85) 214/277 0.75 (0.471.21)
Daily 37/33 1.25 (0.732.14) 28/21 1.64 (0.863.15) 9/12 0.73 (0.271.97)
p for linear trend 0.33 0.05 0.28
Light green/orange leafy veg-
etables
Less than weekly 66/81 1.0 46/55 1.0 20/26 1.0
Weekly 764/827 1.07 (0.751.52) 552/563 1.13 (0.741.72) 212/264 0.92 (0.491.75)
Daily 105/124 0.98 (0.631.50) 70/81 0.97 (0.581.63) 35/43 0.95 (0.442.05)
p for linear trend 0.79 0.76 0.95
Green beans/peas
Less than weekly 720/737 1.0 504/504 1.0 216/233 1.0
Weekly 34/43 0.80 (0.501.28) 23/28 0.84 (0.471.50) 11/15 0.69 (0.301.56)
Daily 181/252 0.77 (0.620.99) 141/167 0.90 (0.691.17) 40/85 0.51 (0.330.79)
p for linear trend 0.02 0.38 0.002
Gourds
Less than weekly 17/18 0.91 (0.451.84) 15/12 1.17 (0.522.64) 2/6 0.38 (0.071.99)
Weekly 301/309 1.0 217/216 1.0 84/93 1.0
Daily 617/705 0.89 (0.731.09) 436/471 0.93 (0.741.18) 181/243 0.81 (0.561.18)
p for linear trend 0.33 0.45 0.53
Green/red peppers
Less than weekly 332/364 1.0 233/253 1.0 99/111 1.0
Weekly 586/641 0.89 (0.731.09) 422/431 0.95 (0.761.21) 164/210 0.75 (0.521.08)
Daily 17/27 0.63 (0.331.20) 13/15 0.84 (0.381.86) 4/12 0.34 (0.101.10)
p for linear trend 0.13 0.62 0.04
Tomatoes
Less than weekly 62/64 1.0 49/46 1.0 13/18 1.0
Weekly 526/524 1.04 (0.701.52) 381/375 0.97 (0.621.52) 145/149 1.18 (0.542.57)
Daily 347/444 0.82 (0.551.20) 238/278 0.86 (0.541.35) 109/166 0.73 (0.331.61)
p for linear trend 0.03 0.29 0.02
All fresh vegetables (times/year)
1st quartile (539) 236/261 1.0 174/193 1.0 62/68 1.0
2nd quartile (539694) 263/256 1.12 (0.871.44) 180/170 1.14 (0.841.54) 83/86 1.05 (0.661.69)
3rd quartile (695865) 231/258 0.95 (0.741.24) 166/173 1.02 (0.751.39) 65/85 0.80 (0.491.31)
4th quartile (866) 205/257 0.85 (0.651.10) 148/163 0.97 (0.711.32) 57/94 0.63 (0.381.04)
p for linear trend 0.13 0.71 0.03
Oranges/tangerines
Less than monthly 167/114 1.0 127/82 1.0 40/32 1.0
Monthly 279/275 0.71 (0.530.96) 195/201 0.64 (0.450.91) 84/74 0.95 (0.531.71)
Weekly 375/488 0.55 (0.410.73) 269/323 0.57 (0.410.80) 106/165 0.51 (0.280.89)
Daily 114/155 0.55 (0.380.78) 77/93 0.61 (0.400.94) 37/62 0.45 (0.240.94)
p for linear trend 0.001 0.008 0.001
1
See Appendix for denitions of various food groups; intake frequency was categorized as follows: less than monthly, 1/month; monthly,
1/month to 1/week; weekly, 1/week to 1/day; daily, 1/day.
2
Ca/Co, number of cases/number of controls.
3
OR, odds ratio, adjusted for age,
gender (for total only), level of education, cigarette smoking, exposure to smoke from heated rapeseed oil and burning coal during cooking,
occupational exposure to chemical fumes and history of chronic ear and nose condition (see text for more detailed description of confounding
variables); CI, condence interval.
361
DIETAND NPC
et al., 1990). In Taiwan, fermented bean products and smoked meat
have been found to be NPC-associated foods (Chen et al., 1988). In
Malaysian Chinese, salted eggs and various salted vegetables
showed independent effects on NPC risk (Armstrong et al., 1998).
Intake of preserved foods also has been shown to be associated
with NPC risk among Arabs of North Africa, another high-risk
population for NPC (Jeannel et al., 1990). As with Chinese salted
sh, carcinogenic nitrosamines and other genotoxic substances
have been detected in samples of many of these NPC-associated
preserved foods (Yu and Henderson, 1996). The ndings of our
present study add to the evidence that a variety of preserved foods
other than Chinese salted sh play a role in the development of
NPC.
Our data demonstrate that NPC patients had ingested signi-
cantly fewer oranges/tangerines than control subjects. The protec-
tive effect of this fruit could not be explained by the differing
consumption pattern of salted sh and other preserved foods
between cases and controls. Oranges and tangerines are a rich
source of vitamin C in the diet. It is known that vitamin C can block
nitrosamine formation in vivo (Raineri and Weisburger, 1975),
thereby offering a biological rationale for the observed protective
effect of oranges/tangerines in NPC development. In a case-control
study, Farrow et al. (1998) reported that high intake of vitamin C
was associated with a signicant, dose-dependent reduction in risk
of NPC among US blacks and whites.
In summary, our present results show that exposure to salted sh
and other preserved foods is a risk factor for NPC in Shanghai,
China. In contrast, high intake of oranges/tangerines confers a
signicant effect on reduction in risk of NPC.
ACKNOWLEDGEMENT
We thank the staff of the Shanghai Cancer Registry for their
assistance in identifying and verifying cancer diagnoses.
REFERENCES
ARMSTRONG, R.W., ARMSTRONG, M.J., YU, M.C. and HENDERSON, B.E.,
Salted sh and inhalants as risk factors for nasopharyngeal carcinoma in
Malaysian Chinese. Cancer Res., 43, 29672970 (1983).
ARMSTRONG, R.W., IMREY, P.B., LYE, M.S., ARMSTRONG, M.J., YU, M.C. and
SANI, S., Nasopharyngeal carcinoma in Malaysian Chinese: salted sh and
other dietary exposures. Int. J. Cancer, 77, 228235 (1998).
BRESLOW, N.E. and DAY, N.E., Statistical methods in cancer research, Vol.
1. The analysis of case-control studies, IARC Sci. Publ. 32, IARC, Lyon
(1980).
CHEN, C.-J., WANG, Y.-F., SHIEH, T., CHEN, J.-Y. and LIU, M.-Y., Multifacto-
rial etiology of nasopharyngeal carcinoma. Epstein-Barr virus, familial
tendency and environmental cofactors. In: G.T. Wolf and T.E. Garey (eds.),
Head and neck oncology research, pp. 469476, Kugler, Amsterdam
(1988).
FARROW, D.C., VAUGHAN, T.L., BERWICK, M., LYNCH, C.F., SWANSON, G.M.
and LYON, J.L., Diet and nasopharyngeal cancer in a low-risk population.
Int. J. Cancer, 78, 675679 (1998).
GESER, A., CHARNEY, N., DAY, N.E., HO, H.C. and DE THE, G., Environmen-
tal factors in the etiology of nasopharyngeal carcinoma: report on a
case-control study in Hong Kong, IARC Sci. Publ. 20, pp. 213229, IARC,
Lyon (1978).
HENDERSON, B.E., LOUIE, E., JING, J.S., BUELL, P. and GARDNER, M.B., Risk
factors associated with nasopharyngeal carcinoma. N. Engl. J. Med., 295,
11011106 (1976).
HUANG, D.P., SAW, D., TEOH, T.B. and HO, J.H.C., Carcinoma of the nasal
and paranasal regions in rats fed Cantonese salted marine sh, IARC Sci.
Publ. 20, pp. 315328, IARC, Lyon (1978).
JEANNEL, D., HUBERT, A., DE VATHAIRE, F., ELLOUZ, R., CAMOUN, M., BEN
SALEM, M., SANCHO-GARNIER, H. and DE THE, G., Diet, living conditions
and nasopharyngeal carcinoma in Tunisiaa case-control study. Int. J.
Cancer, 46, 421425 (1990).
NATIONAL CANCER CONTROL OFFICE AND NANJING INSTITUTE OF GEOGRA-
PHY, Atlas of cancer mortality in the Peoples Republic of China, pp. 8086,
Shanghai China Press, Shanghai (1979).
NING, J.-P., YU, M.C., WANG, Q.-S. and HENDERSON, B.E., Consumption of
salted sh and other risk factors for nasopharyngeal carcinoma (NPC) in
Tianjin, a low-risk region for NPC in the Peoples Republic of China. J. nat.
Cancer Inst., 82, 291296 (1990).
PARKIN, D.M., MUIR, C.S., WHELAN, S.L., GAO, Y.-T., FERLAY, J. and
POWELL, J. (eds.), Cancer incidence in ve continents, Vol. 6. IARC
Sci.Publ. 120, IARC, Lyon (1992).
RAINERI, R. and WEISBURGER, J.H., Reduction of gastric carcinogens with
ascorbic acid. Ann. N.Y. Acad. Sci., 258, 181189 (1975).
YU, M.C. and HENDERSON, B.E., Nasopharyngeal cancer. In: D. Schotten-
feld and J. Fraumeni, Jr. (eds.), Cancer epidemiology and prevention, pp.
603618, Oxford University Press, NewYork (1996).
YU, M.C., HO, J.H.C., LAI, S.-H. and HENDERSON, B.E., Cantonese-style
salted sh as a cause of nasopharyngeal carcinomareport of a case-
control study in Hong Kong. Cancer Res., 46, 956961 (1986).
YU, M.C., HO, J.H.C., ROSS, R.K. and HENDERSON, B.E., Nasopharyngeal
carcinoma in Chinese: salted sh or inhaled smoke? Prev. Med., 10, 1524
(1981).
YU, M.C., HUANG, T.B. and HENDERSON, B.E., Diet and nasopharyngeal
carcinoma: a case-control study in Guangzhou, China. Int. J. Cancer, 43,
10771082 (1989a).
YU, M.C., MO, C.C., CHONG, W.X., YEH, F.-S. and HENDERSON, B.E.,
Preserved foods and nasopharyngeal carcinoma: a case-control study in
Guangxi, China. Cancer Res., 48, 19541959 (1988).
YU, M.C., NICHOLS, P.W., ZOU, X.-N., ESTES, J. and HENDERSON, B.E.,
Induction of malignant nasal cavity tumours in Wistar rats fed Chinese
salted sh. Brit. J. Cancer, 60, 198201 (1989b).
YUAN, J.-M., WANG, X.-L., XIANG, Y.-B., GAO, Y.-T., ROSS, R.K. and YU,
M.C., Non-dietary risk factors for nasopharyngeal carcinoma in Shanghai,
China. Int. J. Cancer, 85, 364369 (2000).
TABLE IV RISK OF NPCACCORDING TO COMBINED LEVELS OF INTAKE OF PRESERVED FOODS AND
ORANGES/TANGERINES IN SHANGHAI, CHINA
1
Intake of oranges/tangerines
3rd tertile (high) 2nd tertile 1st tertile (low)
Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI) Ca/Co
2
OR
3
(95% CI)
Intake of all preserved foods
1st tertile (low) 44/113 1.0 90/109 1.99 (1.263.15) 105/120 2.15 (1.383.37)
2nd tertile 83/103 1.88 (1.182.99) 98/133 1.73 (1.112.71) 135/114 2.76 (1.784.29)
3rd tertile (high) 74/102 1.82 (1.142.92) 120/111 2.36 (1.513.69) 186/127 3.19 (2.084.91)
1
See Appendix for denitions of various food groups; for oranges/tangerines: 1st tertile, 13 times/year;
2nd tertile, 13 to 25 times/year; 3rd tertile, 26 times/year; for all preserved foods: 1st tertile, 111
times/year; 2nd tertile, 111 to 259 times/year; 3rd tertile, 260 times/year.
2
Ca/Co, number of
cases/number of controls.
3
OR, odds ratio, adjusted for age, gender, level of education, cigarette smoking,
exposure to smoke from heated rapeseed oil and burning coal during cooking, occupational exposure to
chemical fumes and history of chronic ear and nose condition (see text for more detailed description of
confounding variables); CI, condence interval.
362
YUAN ET AL.
ZHENG, X., LUO, Y., CHRISTENSSON, B. and DRETTNER, B., Induction of nasal
and nasopharyngeal tumours in Sprague-Dawley rats fed with Chinese
salted sh. Acta Otolaryngol. (Stockholm), 114, 98104 (1994).
ZHENG, Y.M., TUPPIN, P., HUBERT, A., JEANNEL, D., PAN, Y.J., ZENG, Y. and
DE THE, G., Environmental and dietary risk factors for nasopharyngeal
carcinoma: a case-control study in Zangwu County, Guangxi, China. Brit. J.
Cancer, 69, 508514 (1994).
APPENDIX
Food items included in the study questionnaire:
1. Protein-containing preserved foods: harm yu (salted sh),
gone yu (dried sh), harm ha (salted shrimp paste), hoi sin cheung
(salted sh paste), lup yoke (ham), lup arp (cured duck), harm ju
yoke (salted pork), harm dann (salted egg), pay dann (fermented
black egg), fu yu (fermented bean curd), tau ban cheung (fermented
bean paste), tau si (salted black bean).
2. Preserved vegetables/fruits: mui choi (salted mustard green),
harm choi (salted bok choi), harm wou sun (salted stem vegetable),
cheung wou sun (pickled stem vegetable), cheung wong
qua (pickled cucumber), chung choi (salted root), zha choi
(preserved root), chan pai mui (dried plum), gar ink gee (preserved
plum), gone hong zhao (dried red date), bei guo (dried gingko
seed).
3. Fermented sauces: see yao (soy sauce), yu lo (sh sauce), ho
yao (oyster sauce).
4. Fresh vegetables/fruits: bok choi (Chinese white cabbage),
other dark green leafy vegetables, light green/yellow leafy veg-
etables, fresh beans/peas, fresh gourds, tomatoes, red/green pep-
pers, oranges/tangerines.
5. Other foods: fresh animal organs, fresh eggs, fresh sh, fresh
cows milk.
363
DIETAND NPC