Novel epidemiologic evidence for the association between fermented fish sauce and esophageal cancer in South China
Article first published online: 20 MAR 2002
Copyright © 2002 Wiley-Liss, Inc.
International Journal of Cancer
Volume 99, Issue 3, pages 424–426, 20 May 2002
How to Cite
Ke, L., Yu, P. and Zhang, Z. X. (2002), Novel epidemiologic evidence for the association between fermented fish sauce and esophageal cancer in South China. Int. J. Cancer, 99: 424–426. doi: 10.1002/ijc.10293
- Issue published online: 25 APR 2002
- Article first published online: 20 MAR 2002
- Manuscript Accepted: 18 DEC 2001
- Manuscript Revised: 7 DEC 2001
- Manuscript Received: 23 AUG 2001
- The Ministry of Education of PRC
- Li Ka Shing Foundation
- Office of Hygiene of Guangdong Province
- case-control studies;
- risk factors;
- fermented fish sauce
Previous studies have suggested that fermented fish sauce is related to an increased risk for nasopharyngeal, thyroid and gastric cancers and has suspicious carcinogenic and promoting effects in the laboratory, but these reports have not investigated the association between this agent and esophageal cancer in population. Therefore, we investigated the relationship between esophageal cancer and consumption of fermented fish sauce, alcohol and tobacco after adjusting other risk factors using data from a large hospital-based case-control study in Chaoshan area of China. The subjects for analysis included 1,248 cases (median age 58.5 [range 29–82] years, 936 males, 312 females, some 50% with the habit of fermented fish sauce eating) with squamous cell carcinoma of the esophagus and the same amount of controls matched by sex and age. A significant increase in risk (OR 3.21; 95% CI 2.45–4.19) for eating fermented fish sauce compared to not eating and a significant dose-response relationship was found with the consumption of fermented fish sauce (p for trend < 0.001). In comparison based on the binary variables, the OR for the subjects with fermented fish sauce eating and current smoking was 15.5 with a 95% CI of 8.14–29.3, relative to those exposed to neither habit. The joint effect of fermented fish sauce (F) and smoking (S) was more than additive (F * S > F + S: 15.5 > 1 + (5.6 − 1) + (6.6 − 1)), but the effect for fermented fish sauce and alcohol was not found. It is concluded that risks of esophageal cancer in the population were substantially associated with fermented fish sauce. Further epidemiologic and experimental study are required to find a biologic causal relationship between them. © 2002 Wiley-Liss, Inc.
Chaoshan area of Guangdong in South China is a high risk 1 for esophageal cancer, with the mortality rates of esophageal cancer ranging from 100 per 105 in Nanao, an island and islet county, to 28 per 105 in the city of Shantou. The very high risk of esophageal cancer may be related to a particular cause in this area. Yulu (transliterate, known by the folk name of harm ha in Philippines) is a salty liquid condiment made by fermenting small fishes and shrimps in brine, so it may be called fermented fish sauce, fish sauce or fish juice. There are a substantial number of Yulu eaters in the local population and some areas in Southeast Asia, thus it is essential for us to study whether fermented fish sauce is a risk factor of esophageal cancer. It have been shown that fermented fish sauce has suspicious carcinogenic and promoting effects in experimental animals.1, 2, 3, 4, 5, 6 Some epidemiologic investigations have suggested that fermented fish sauce related to an increased risk for nasopharyngeal cancer in Guangzhou, China,7 thyroid cancer in Hawaii8 and gastric cancer in Fujian, China.9 So far as we know, however, epidemiologic evidence on a relationship between fermented fish sauce and esophageal cancer remains absent. Here we analyzed the relationship among esophageal cancer and consumption of fermented fish sauce, alcohol and tobacco after adjusting other risk factors using data from a large hospital-based case-control study.
MATERIAL AND METHODS
Patients from the 2 local hospitals and the 3 university hospital in-patient departments were included if they had a histologic diagnosis of esophageal squamous cell carcinoma from October 1997 to June 2000 and lived in the country for at least 10 years. For each case, there was 1 control, who also lived in the country for at least 10 years, was matched by age (±3 years), sex and same hospital and had not had a diagnosis of tobacco and/or alcohol-related diseases. The same questionnaires for cases and controls were completed by 8 trained medical students. Information was collected on occupation, socioeconomic status (income and education), alcohol drinking (dose, expressed in milliliters of alcohol, duration and type of alcoholic beverage) and tobacco smoking (quantity, duration, intensity and cessation periods). Smoking and dietary habits were assessed in the period just before 10 years before admission to the hospital. A food frequency questionnaire was used to assess the consumption of the various food items. The 12 food items used in the study questionnaire were sowbelly, kipper, pickled vegetable, smoked food, fermented fish sauce, vegetables, fresh fruits, fresh meat, eggs, cereals, fats and dairy products.
An unconditional analysis was done because it is generally accepted that in a matched study, this analysis can be performed if appropriate adjustment is made for matching variables, which in our study were sex, age and hospital. Moreover, the results obtained with the conditional analysis were practically identical to those obtained with the unconditional one. Both sexes were combined in statistical analysis.
General status and analysis for univariable
A total of 1,248 cases (median age 58.5 [range 29–82] years, 936 males, 312 females, some 50% with the habit of Yulu eating) and the same amount of controls were included in our study. Their distribution by age, sex, residence, income and some risk factors is given in Table I. Cases, especially males, are more often farmers and fishermen, from high-incidence areas and have lower economic status than controls.
|Farmer and fisherman||58.3||7.1||61.5||9.3||59.1||7.7|
|Income (in RMB)|
Distribution of fermented fish sauce in region and occupation
Since region and occupation are potentially strong confounders, a relationship between them and fermented fish sauce were analyzed in Table II. High-incidence areas, especially among cases, contained more Yulu eaters, but there were not significant differences among occupation groups.
|Region and occupation||Prevalence rate (%) of fermented fish sauce eating|
|Farmer and Fisherman||45.8||17.7||42.6|
Dose-response analysis for fermented fish sauce
Unconditional logistic regression analysis with stepwise forward selection (probability for entry 0.05; probability for removal 0.10) of variables showed that those with the risk effect were older, male, rural inhabitants, farmers or fishermen, had lower income and consumed tobacco, alcohol, fermented fish sauce, pickled vegetable and sowbelly. Those with the protective effect were fresh fruit and cooked vegetable eaters (but there was a lack of a dose-response relationship). No significant associations with the other food items were observed. The entry and exit criteria used in the stepwise model are lower than most recommend (0.15 for entry and 0.2 for exit) in order to select the covariates significantly associated with risk.
An analysis for fermented fish sauce after adjustment for age, sex, region, occupation, income, alcohol, tobacco, pickled vegetable, sowbelly, fruit and vegetable were carried out. Table III shows that a significant increase in risk and a significant dose-response relationship was found with the consumption of fermented fish sauce, but the risks between 2–3 times/week and >3 times/week remained almost constant.
|Case||Control||OR (95% CI1)|
|Fermented fish sauce|
|2–3 times/week||414||336||2.09 (1.59–2.75)|
|>3 times/week||76||8||14.29 (5.88–35.4)|
|Trend test (v = 1)||x2 = 105.6; p < 0.001|
Joint effect of fermented fish sauce and tobacco or alcohol
For further analyses, cigarette smoking was divided into nonsmoking and current smoking (1 cigarette/day or more) after the ex-smokers, whose smoking status might be uncertain, were excluded. Alcohol-drinking groups were divided into either daily or not-daily (3 times/week or less) types. The joint effect of alcohol and smoking was more than additive (data not show). In Table IV, the OR for fermented fish sauce and current smoking was 15.5 with a 95% CI of 8.1–29.3, relative to men and women exposed to neither habit. Thus, the joint effect of both was more than additive (e.g., 15.5 > 1 + (5.6 − 1) + (6.6 − 1)). However, the effect for fermented fish sauce and alcohol was not found (Table V).
|Fermented fish sauce||Tobacco (current)||No. cases||No. controls||OR (95% CI)|
|Fermented fish sauce||Alcohol (daily)||No. cases||No. controls||OR (95% CI)|
Our study was designed to study the association between fermented fish sauce and esophageal cancer. A strong association with a clear dose-response relationship was observed with the fermented fish sauce consumption. The OR for those eating daily was 15.8 (95% CI, 7.0–35.4), after adjusting for other potential factors. There are 2 possible mechanisms through which fermented fish sauce could increase the risk of esophageal cancer. First, the condiment extract may contain carcinogenic substances. This possibility was raised in experimental study carried out in China.3, 4, 5, 6 Direct mutagenicity and high contents of N-nitrosamide were detected when fermented fish sauce samples were nitrosated at pH 2.0. When the nitrosated fermented fish sauce extract was given to newborn rats by gavage, dysplasia and adenocarcinoma were induced in the glandular stomach.3 This genotoxicity was partly ascribed to the formation of nitrite-derived arene diazonium cations that were characterized by a coupling reaction with N-ethyl-1-naphthylamine and thin-layer chromatography.4 Further studies have found formation of N-(Nitrosomethyl)urea (NMU) in stomachs of experimental pigs and human volunteers given fish sauce in vivo,6 It has been confirmed that the carcinogenic NMU was derived by nitrosation of fermented fish sauce.5 The laboratory observations suggest that fermented fish sauce itself may not contain specific carcinogens but substantive carcinogenic N-nitroso compounds may be synthesized after it comes into stomach. Second, fermented fish sauce may contain promoting substances. The experimental results show that the homemade fermented fish sauce proved a distinct promoting effect on the process of cocarcinogenesis initiated by NSEE in the forestomach of mice,2 while the market fermented fish sauce has no significant promoting effect on the forestomach epithelium of mice and the esophageal epithelium of rats. Homemade fermented fish sauce sample contained tumour promoterlike substances, as measured by expression of certain EBV genes in Raji cells.4
Several epidemiologic studies point toward a possible effect of fermented fish sauce on cancers of some sites such as rhino- pharynx, thyroid and stomach,7, 8, 9 but not including esophagus. In particular, a study from Fujian, China,9 suggests that inhabitants of high-risk areas for gastric cancer eat large quantities of fermented fish sauce.
However, the explanation of the association with esophageal cancer remains obscure. There is no obvious reason why they should cause cancer in esophagus. The fermented fish sauce comes into contact first and in highest concentration with the mucosa of the mouth and esophagus, but contact is transient and of very short duration, especially in the esophagus. Therefore, as might be expected, fermented fish sauce has a profound effect on the metabolism of many hormones, metabolic intermediates, toxic chemicals and carcinogens and it is as a result of these interactions that fermented fish sauce may enhance the effect of environmental carcinogens such as those contained in tobacco tar. Our results on the joint effect between fermented fish sauce and tobacco support this possibility.
We thank Mr. Y.H. Lin, Mr. H. Ge and Ms. M.X. Hong for their help and Dr. N.E. Day for critical comments.