Abstract
AIM: To study the main environmental and lifestyle factors that account for the regional differences in esophageal cancer (EC) risk in low- and high-risk areas of Jiangsu Province, China.
METHODS: Since 2003, a population-based case-control study has been conducted simultaneously in low-risk (Ganyu County) and high-risk (Dafeng County) areas of Jiangsu Province, China. Using identical protocols and pre-tested standardized questionnaire, following written informed consent, eligible subjects were inquired about their detail information on potential determinants of EC, including demographic information, socio-economic status, living conditions, disease history, family cancer history, smoking, alcohol drinking, dietary habits, frequency, amount of food intake, etc. Conditional logistic regression with maximum likelihood estimation was used to obtain Odds ratio (OR) and 95 % confidence interval (95% CI), after adjustment for potential confounders.
RESULTS: In the preliminary analysis of the ongoing study, we recruited 291 pairs of cases and controls in Dafeng and 240 pairs of cases and controls in Ganyu, respectively. In both low-risk and high-risk areas, EC was inversely associated with socio-economic status, such as level of education, past economic status and body mass index. However, this disease was more frequent among those who had a family history of cancer or encountered misfortune in the past 10 years. EC was also more frequent among smokers, alcohol drinkers and fast eaters. Furthermore, there was a geographic variation of the associations between smoking, alcohol drinking and EC risk despite the similar prevalence of these risk factors in both low-risk and high-risk areas. The dose-response relationship of smoking and smoking related variables, such as age of the first smoking, duration and amount were apparent only in high-risk areas. On the contrary, a dose-response relationship on the effect of alcohol drinking on EC was observed only in low-risk areas.
CONCLUSION: The environmental risk factors, together with genetic factors and gene-environmental interactions might be the main reason for this high-risk gradient in Jiangsu Province, China.
Keywords: Esophageal cacner, Case-control study, Smoking, Alcohol drinking, Dietary factors
INTRODUCTION
Esophageal cancer (EC) is the sixth most common cause of cancer mortality worldwide. The incidence of this disease shows a striking geographic variation in the world; a 20-fold variation is observed between high-risk China and low-risk western Africa[1]. Jiangsu Province, south-eastern China is one of the highest EC incidence areas with a mortality rate of 30.0/100 000 between 1990-1992, which was significantly higher than the national average of 17.0/100 000[2,3]. EC has been the third leading cause of cancer mortality in Jiangsu Province since the 1970s[3]. Although the mortality rates of EC are high in most counties of Jiangsu, national surveys conducted in the 1970s and 1990s have shown that rates differ considerably between different counties within the province, despite their similar geographic characteristics and socioeconomic status[3,4].
Comprehensive studies on the etiology and
carcinogenesis of EC in high-risk areas have been carried out during the past decades. Epidemiological evidence suggests that the independent risk factors, tobacco smoking and alcohol drinking, are strongly associated with EC risk and have approximately multiplicative joint effects[5-7]. Dietary factors were also found to play an important role in the development of EC. Increased risk of EC was found to be associated with low intake of raw vegetables and fresh fruits, a deficiency in vitamins or protective antioxidants (e.g. Vitamin C and E, β-carotene, and selenium), high intake of carcinogens (frequent consumption of pickled vegetables and fungi toxins) and thermal injuries (fast eating speed for hot drinks and soups)[8-10]. While the contributory factors of EC are the high consumption of tobacco and alcoholic beverages in Western countries, the causative factors of EC in high-risk areas of China are nutrition deficiency, N-nitrosamines, fungi toxins and genetic factors[11].
Although numerous epidemiological studies have been conducted to explore the associations between environmental, lifestyle, dietary factors and the risk of esophageal cancer, few studies have been conducted to compare the association between risk factors and EC in apparently similar areas with a high risk gradient. Thus, a population-based case-control study has been conducted since 2003 in both low and high-risk areas of Jiangsu Province, China to study main environmental and lifestyle factors that account for regional differences in EC risk.
This paper reports the preliminary results on the independent and joint effects of smoking, alcohol drinking and dietary factors on EC risk and compares their associations with EC in both high-risk and low-risk areas.
MATERIALS AND METHODS
Study area
This study has been conducted in Dafeng and Ganyu counties since late 2003. Both counties are less developed coastal areas in northern Jiangsu Province, China, with 0.7 million and 1.1 million inhabitants, respectively. Dafeng is a high cancer incidence area and has a much higher mortality rate of EC than Ganyu. From 1996 to 2002, the yearly average age-adjusted mortality rate of EC in Dafeng was 36/100 000, whereas Ganyu had a considerably lower age-adjusted EC mortality rate of 24/100 000 during the same period (P < 0.01)[12].
Selection of cases and controls
Cases Newly diagnosed patients with primary esophageal cancer were recruited using data from regional cancer registry agencies. The cancer registry agencies in both counties were established in the late 1990s and are connected to the local Center for Disease Control and Prevention (CDC). All cases were coded according to the International Classification of Diseases, tenth revision (ICD-10, code C15.0 to C15.9). Secondary and recurrent cancers were excluded. All cases were restricted to local inhabitants of the two counties who have lived in either area for at least 5 years. In 2004, 45% and 72% of all newly registered EC cases were recruited and interviewed in Dafeng and Ganyu, respectively. The comparatively low response rate in Dafeng was partly due to the low involvement of local hospitals during the beginning of the study. A small number of cases were also unwilling to participate. Presently, the response rate in Dafeng is much higher. A system of rapid case recognition was used in the study. All regional hospitals were required by the local health authorities to report new EC patients shortly after diagnosis. As the cancer registry agencies are connected to the local CDC, the field investigators from the local CDC were able to identify and interview most patients within one month after their diagnosis. Of all the EC cases in Dafeng, 46% were histologically confirmed, 40% and 13% were diagnosed by endoscopy and radiology, respectively. In Ganyu, 30% of EC cases were histologically confirmed, 50% and 16% were diagnosed by endoscopy and radiology, respectively.
Controls Cases and controls were individually matched and derived from the same county. The criteria for the eligibility of controls were established as: controls had to be the same gender and within 5 years of age as the case, had to have lived in the area for at least 5 years, and had to have had good physical and mental health to answer questions reliably. Controls were randomly selected by a computer from the demographic database of the general population in the county police station. Local interviewers were responsible for locating and interviewing controls. If a selected control refused to participate, a replacement was found using the same recruitment criteria. The response rate of controls in both Dafeng and Ganyu was around 70%. Till March 2005, more than 400 EC cases were recruited in each county. However, the control recruitment rate is slower than the rate of case identification; thus only 291 and 240 pairs of cases and their matched controls were used in this analysis.
Data collection
Identical protocols and pre-tested standardized questionnaires were used. Data collection included a written informed consent, a face-to-face interview, a physical examination, and a 5 ml blood sample taken by professional interviewers from the local CDC in both counties. The questionnaire elicited detailed information on potential determinants of EC, including demographic information, socio-economic status, living conditions, disease history, family cancer history (any malignant neoplasm in first-degree relatives), smoking, alcohol drinking, dietary habits, and frequency and amount of food intake.
In our study, never-smokers were defined as having smoked fewer than 100 cigarettes in their lifetime. Current smokers and drinkers were defined as those who had the habit during the time of interview or those who stopped the habit because of health problems within one year of the date of interview. The dietary questionnaire used in this study included 90 food items. For each food item, the amount and frequency of consumption over the past year were inquired. For cases, the amount and frequency of consumption referred to the year prior to the onset of the disease. In the final analysis, foods were categorized into several major groups: staple foods, preserved foods, meat, fish, eggs, soybean, and fruits and vegetables.
An anthropometric measurement and physical examination also took place at the time of interview to evaluate the subject’s health status. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared (i.e., kg/m2). BMI was grouped into quartiles according to the Chinese national standard (underweight: <18.5, normal: 18.5-23.9, overweight: 24~27.9, obesity: ≥28)[13].
Statistical analysis
Data were double entered using Epidata 2.1b and cleaned and analyzed using SAS v8.2. Chi-square and Student t-tests were used to compare the distribution of relevant factors among the control groups between the two counties. Conditional logistic regression with maximum likelihood estimation of parameters was applied for both univariate and multivariate analyses. This was done by transforming each matched pair into a single observation, where the explanatory variable value was the difference between the corresponding values for the case-control pair [14]. Continuous variables such as income level and amount of food intake were divided into quartiles based on the frequency distribution among control groups.
The strength of the association was quantified as odds ratio (OR) obtained from conditional logistic regression. Statistical significance was set at 0.05, and accordingly, 95% confidence interval (CI) around the OR was used to address precision.
RESULTS
Subject characteristics
There were 291 pairs of cases and controls (200 male and 91 female pairs) in Dafeng and 240 pairs (181 male and 59 female) in Ganyu, respectively. By design, cases and controls had similar distributions in terms of gender and age in both two counties (Table 1). The differences in the distribution of the above-mentioned variables between the two counties were examined by comparing the two control groups. The proportion of patients who were older than 70 years of age and the proportion of illiteracy in Ganyu were higher than that of Dafeng (P = 0.002 and P < 0.001, respectively). There was also a geographic variation between the two counties between the proportion of BMI (P = 0.014), occurrences of misfortunes such as fire disasters, lost of family members, divorces, etc in the past 10 years (P = 0.008) and family history of cancer (P < 0.001). On the other hand, there was no significant difference in past economic status, ever-smoking and ever-alcohol drinking habits between the two counties.
Table 1.
Characteristics among EC cases and controls in Dafeng and Ganyu 1 n (%)
| Characteristics | Dafeng (high-risk) | Ganyu (low-risk) | P value2 | |||
| Case (n=291) | Control (n=291) | Control (n=240) | Case (n=240) | |||
| Gender: Male | 200 (68.7) | 200 (68.7) | 181 (75.4) | 181 (75.4) | 0.088 | |
| Female | 91 (31.3) | 91 (31.3) | 59 (24.6) | 59 (24.6) | ||
| Age (yr) Mean±SD | 64.8 ± 8.6 | 64.6 ± 8.9 | 65.4 ± 10.3 | 65.6 ± 10.4 | 0.002 | |
| <50 | 14 (4.8) | 17 (5.8 ) | 19 (7.9) | 17 (7.1) | ||
| 50-59 | 61 (30.0) | 59 (20.3) | 48 (20.0) | 51 (21.3) | ||
| 60-69 | 137(47.1) | 137 (47.1) | 78 (32.5) | 76 (31.7) | ||
| 70-79 | 71 (24.4) | 69 (23.7) | 77 (32.1) | 78 (32.5) | ||
| ≥80 | 8 (2.8) | 9 (3.1) | 18 (7.5) | 18 (7.5) | ||
| Level of education | ||||||
| Illiterate | 156 (53.6) | 130 (44.7) | 138 (57.7) | 164 (68.6) | <0.001 | |
| Primary school | 95 (32.7) | 119 (40.9) | 63 (26.4) | 54 (22.6) | ||
| Secondary school & above | 40 (13.7) | 42 (14.4) | 38 (15.9) | 21 (8.8 ) | ||
| Past economic status(By separate cut-off points) | ||||||
| Median (CNY/yr) | 1250 | 1500 | 1000 | 775 | 0.235 | |
| 1 (lowest) | 97 (33.5) | 55 (19.0) | 38 (16.2) | 47 (20.3) | ||
| 2 | 68 (23.5) | 64 (22.1) | 71 (30.2) | 87 (37.5) | ||
| 3 | 73 (25.2) | 96 (33.1) | 71 (30.2) | 59 (25.4) | ||
| 4 (highest) | 52 (17.9) | 75 (25.9) | 55 (23.4) | 39 (16.8) | ||
| Smoking status3 | ||||||
| Neversmoker | 92 (31.6) | 122 (41.9) | 95 (39.6) | 82 (34.2) | 0.067 | |
| Former-smoker | 71 (24.4) | 64 (22.0) | 19 (7.9) | 17 (7.1) | ||
| Current smoker | 128 (44.0) | 105 (36.1) | 126 (52.5) | 141 (58.7) | ||
| Alcohol drinking status4 | ||||||
| Never drinker | 175 (60.1) | 181 (62.2) | 143 (59.6) | 131 (54.6) | 0.076 | |
| Former drinker | 5 (1.7) | 7 (2.4) | 7 (2.9) | 7 (2.9) | ||
| Current drinker | 111 (38.1) | 103 (35.4) | 90 (37.5) | 102 (42.5) | ||
| Encountered misfortune in past10 yr: | ||||||
| No | 54 (18.6) | 41 (14.2) | 33 (14.0) | 54 (23.3) | 0.008 | |
| Yes | 237 (81.4) | 248 (85.8) | 203 (86.0) | 178 (76.7) | ||
| History of family cancer | ||||||
| No | 112 (38.5) | 86 (29.6) | 16 (6.7) | 29 (12.1) | <0.001 | |
| Yes | 179 (61.5) | 205 (70.5) | 224 (93.3) | 211 (87.9) | ||
| Body mass index | ||||||
| <18.5 | 60 (20.8) | 27 (9.3) | 18 (7.6) | 31 (13.4) | 0.014 | |
| 18.5-23.9 | 182 (63.0) | 192 (66.2) | 161 (67.9) | 155 (67.1) | ||
| 23.9-27.9 | 34 (11.8) | 60 (20.7 ) | 46 (19.4) | 28 (12.1) | ||
| ≥28 | 13 (4.5) | 11 (3.8) | 12 (5.1) | 17 (7.4) | ||
1 Some strata do not match the total because of missing values;
The P-value for comparing the distribution of factors between the two counties;
Never-smokers and ever-smokers were used for comparing the smoking habits between the two counties;
Because of the few numbers of former drinkers in both two counties, alcohol drinking status was categorized to never-drinkers and ever-drinkers for the comparison between the two counties and following analyses.
Socio-economic status
EC occurred less in higher socio-economic groups characterized by high levels of education and high economic status in both counties (Table 2). On the contrary, low levels of education, low economic status, family history of cancer in first-degree relatives (Dafeng OR: 1.53, Ganyu OR: 2.07), and occurrences of misfortune in the past 10 years (Dafeng OR: 1.26, Ganyu OR: 1.64) increased the risk of developing EC in both areas. In Dafeng, when compared to the lowest quartile (underweight people) of BMI, the second quartile (normal weight people, OR = 0.45) and the third quartile (overweight people, OR = 0.26) significantly showed a reduced risk of EC; whereas the OR increased in the highest quartile (obese people, OR = 0.49). A similar association between BMI and EC risk was also found in Ganyu, although the trend was not significant.
Table 2.
OR1 and 95% CI of socioeconomic status in EC of Dafeng and Ganyu
| Socioeconomic status | Dafeng (high-risk) | Ganyu (low-risk) | |
| Level of education | |||
| Illiterate | 1.00 (Referent) | 1.00 (Referent) | |
| Primary school | 0.54 (0.35-0.84) | 0.58 (0.33-1.03) | |
| Secondary school & above | 0.74 (0.39-1.41) | 0.42 (0.21-0.83) | |
| P value for trend | 0.08 | 0.008 | |
| Past economic status | |||
| 1 (lowest) | 1.00 (Referent) | 1.00 (Referent) | |
| 2 | 0.67 (0.43-1.06) | 1.03 (0.61-1.75) | |
| 3 | 0.44 (0.28-0.68) | 0.76 (0.44-1.31) | |
| 4 (highest) | 0.39 (0.23-0.65) | 0.73 (0.41-1.28) | |
| P value for trend | <0.001 | 0.024 | |
| Encountered misfortune in past 10 yr | |||
| No | 1.00 (Referent) | 1.00 (Referent) | |
| Yes | 1.26 (0.80-1.98) | 1.64 (0.98-2.73) | |
| History of family cancer | |||
| No | 1.00 (Referent) | 1.00 (Referent) | |
| Yes | 1.53 (1.06-2.19) | 2.07 (1.03-4.17) | |
| Body mass index | |||
| <18.5 | 1.00 (Referent) | 1.00 (Referent) | |
| 18.5-23.9 | 0.45 (0.26-0.76) | 0.50 (0.28-0.90) | |
| 23.9-27.9 | 0.26 (0.13-0.50) | 0.36 (0.17-0.75) | |
| ≥28 | 0.49 (0.18-1.33) | 0.80 (0.33-1.98) | |
| P value for trend | 0.002 | 0.376 | |
1 Matched by age and gender, further adjusted for education level and past economic status (quartile).
Tobacco smoking and alcohol drinking
Consistent smoking elevated the risk of developing EC in both counties (Table 3). In Dafeng, former smokers and current smokers have a 1.93- and 2.42- fold higher risk of developing EC than never-smokers. In Ganyu former smoking and current smoking also increased the risk of developing EC (OR = 1.28 and 2.36 respectively). We found in Dafeng that smoking at an earlier age (for trend P = 0.016), long durations of smoking (for trend P = 0.006), and large amounts of cigarettes per day (for trend P = 0.029) were significantly associated with increased EC risk, with an apparent dose-response relationship. However, these associations were not significant in Ganyu.
Table 3.
OR1 and 95% CI of tobacco smoking in EC of Dafeng and Ganyu
| Tobacco smoking | Dafeng (high-risk) | Ganyu (low-risk) | |
| Smoking status2 | |||
| Former smoker | 1.93 (0.91-4.08) | 1.28 (0.28-5.83) | |
| Current smoker | 2.42 (1.28-4.56) | 2.36 (0.89-6.26) | |
| P value for trend | 0.005 | 0.07 | |
| Duration of smoking (yr)2 | |||
| <20 | 2.02 (0.93-4.38) | 1.60 (0.31-7.90) | |
| 20-34 | 2.32 (1.15-4.67) | 2.25 (0.80-6.35) | |
| ≥35 | 1.80 (0.62-5.24) | 0.98 (0.29-3.24) | |
| P value for trend | 0.016 | 0.249 | |
| Duration of smoking (yr)2 | |||
| 1-29 | 1.61 (0.67-3.86) | 1.44 (0.46-4.42) | |
| 30-49 | 2.65 (1.28-5.49) | 2.04 (0.60-6.92) | |
| ≥20 | 2.04 (0.78-5.35) | 1.98 (0.43-9.11) | |
| P value for trend | 0.009 | 0.194 | |
| Amount of smoking (Cig/d)2 | |||
| 1-9 | 1.36 (0.50-3.74) | 1.12 (0.27-4.68) | |
| 10-19 | 2.21 (1.01-4.80) | 1.56 (0.42-5.78) | |
| ≥20 | 2.04 (1.00-4.18) | 0.91 (0.32-2.61) | |
| P value for trend | 0.015 | 0.915 | |
| Total consumption of cigarettes2 | |||
| 1 (lowest) | 1.40 (0.61-3.21) | 0.96 (032-2.82) | |
| 2 | 2.55 (1.06-6.14) | 3.50 (0.37-32.8) | |
| 3 | 1.88 (0.79-4.49) | 1.94 (0.25-14.7) | |
| 4 (highest) | 1.81 (0.57-5.74) | 0.74 (0.19-2.81) | |
| P value for trend | 0.029 | 0.959 | |
1 Matched by age and gender, further adjusted for level of education, past economic status (group) and alcohol drinking;
Never-smokers were used as the reference group.
In Ganyu, subjects who had drunk alcohol tended to have a higher risk of EC (OR = 1.71, 95% CI: 1.02-2.88). Moreover, drinking at an early age (for trend P = 0.012) and long durations of drinking (for trend P = 0.061) showed an increasing association with EC (Table 4). A high consumption of pure ethanol per week, 10 years ago, slightly elevated the risk of EC, but no significant dose-response relationship was found. We did not find any significant association between alcohol drinking and EC in Dafeng, despite its similar alcohol drinking prevalence as Ganyu. The joint effects by smoking and alcohol drinking were also explored in both counties, but no significant interaction was observed either in Dafeng (P = 0.900) or in Ganyu (P = 0.870).
Table 4.
OR1 and 95% CI of alcohol drinking in EC of Dafeng and Ganyu
| Alcohol drinking | Dafeng (high-risk) | Ganyu (low-risk) | |
| Alcohol drinking | |||
| Never | 1.00 (Referent) | 1.00 (Referent) | |
| Ever | 1.01 (0.70-1.46) | 1.71 (1.02-2.88) | |
| P value | 0.964 | 0.043 | |
| Age of first drink2 | |||
| <20 | 0.83 (0.44-1.58) | 2.59 (1.03-6.50) | |
| 20-34 | 1.23 (0.79-1.91) | 1.95 (1.08-3.53) | |
| ≥35 | 0.81 (0.48-1.35) | 1.18 (0.56-2.47) | |
| P value for trend | 0.815 | 0.012 | |
| Duration of drinking (yr)2 | |||
| 1-24 | 0.96 (0.56-1.59) | 1.28 (0.58-2.79) | |
| 25-34 | 0.89 (0.48-1.64) | 1.48 (0.75-2.94) | |
| 35-44 | 1.57 (0.92-2.70) | 1.47 (0.71-3.01) | |
| ≥45 | 0.77 (0.43-1.40) | 1.88 (0.95-3.75) | |
| P value for trend | 0.834 | 0.061 | |
| Alcohol consumption 10 years ago2 (pure ethanol mL/wk) | |||
| 1-249 | 0.87 (0.49-1.54) | 0.79 (0.36-1.74) | |
| 250-499 | 1.06 (0.60-1.89) | 0.61 (0.30-1.25) | |
| 500-749 | 0.97 (0.52-1.79) | 1.63 (0.77-3.43) | |
| ≥750 | 1.10 (0.63-1.93) | 1.27 (0.71-2.28) | |
| P value for trend | 0.74 | 0.223 | |
1 Matched by age and gender, further adjusted for level of education, past economic status (group) and tobacco smoking;
Never-drinkers were used as the reference group.
Dietary factors
After adjusting for potential confounders in both Dafeng and Ganyu, subjects with fast eating speeds showed an increased risk of developing EC (Dafeng: OR=4.01; Ganyu: OR=3.09). On the other hand, high food temperatures, the possibility of being exposed to grain fungi pollution, and frequent intake of fresh garlic did not influence EC risk significantly (Table 5).
Table 5.
OR and 95% CI of dietary habits, food consumption in EC of Dafeng and Ganyu
| Dietary factor | County |
Category |
P value for trend | |||
| 1 (lowest) | 2 | 3 | 4 (highest) | |||
| Food temperature1 | Dafeng | 1 | 0.51 (0.24-1.09) | - | - | 0.08 |
| (1-Normal; 2- Hot) | Ganyu | 1 | 1.14 (0.55-2.41) | - | - | 0.714 |
| Eating speed1 | Dafeng | 1 | 4.01 (1.87-8.62) | - | - | <0.001 |
| (1-Normal; 2-Fast) | Ganyu | 1 | 3.09 (1.24-7.70) | - | - | 0.015 |
| Self reported grain fungi pollution1 | Dafeng | 1 | 2.27 (0.79-6.54) | - | - | 0.131 |
| (1-Likely; 2-Not likely) | Ganyu | 1 | 1.18 (0.45-3.11) | - | - | 0.741 |
| Fresh garlic/wk23 | Dafeng | 1 | 0.64 (0.26-1.60) | - | - | 0.337 |
| Ganyu | 1 | 1.17 (0.57-2.41) | - | - | 0.664 | |
| Staple foods1 | Dafeng | 1 | 0.45 (0.19-1.10) | 0.54 (0.21-1.38) | 0.73 (0.26-2.04) | 0.474 |
| Ganyu | 1 | 0.98 (0.10-9.77) | 0.45 (0.05-4.38) | 0.54 (0.05-4.34) | 0.324 | |
| Meat4 | Dafeng | 1 | 0.73 (0.29-1.85) | 1.66 (0.68-4.10) | 1.93 (0.64-5.77) | 0.16 |
| Ganyu | 1 | 0.54 (0.12-2.42) | 1.17 (0.28-4.92) | 0.65 (0.11-3.67) | 0.305 | |
| Fish and seafood products4 | Dafeng | 1 | 1.14 (0.64-2.03) | 2.11 (1.12-3.96) | 1.91 (1.00-3.64) | 0.023 |
| Ganyu | 1 | 0.98 (0.42-2.28) | 0.64 (0.28-1.44) | 1.04 (0.46-2.33) | 0.794 | |
| Eggs4 | Dafeng | 1 | 0.53 (0.20-1.44) | 1.23 (0.54-2.80) | 1.99 (0.72-5.49) | 0.146 |
| Ganyu | 1 | 0.69 (0.28-1.73) | 0.30 (0.10-1.10) | 0.95 (0.41-2.22) | 0.936 | |
| Soybean45 | Dafeng | 1 | 1.81 (0.88-3.74) | - | - | 0.11 |
| Ganyu | 1 | 1.31 (0.37-4.59) | - | - | 0.677 | |
| Preserved foods4 | Dafeng | 1 | 0.26(0.09-0.75) | 0.49 (0.16-1.46) | 0.94 (0.37-2.36) | 0.635 |
| Ganyu | 1 | 1.05 (0.37-2.97) | 0.56 (0.21-1.48) | 1.21 (0.46-3.20) | 0.932 | |
| Vegetables4 | Dafeng | 1 | 1.26 (0.50-3.16) | 0.94 (0.39-2.30) | 1.37 (0.49-3.83) | 0.72 |
| Ganyu | 1 | 0.34 (0.08-1.54) | 0.80 (0.20-3.18) | 0.76 (0.15-3.72) | 0.889 | |
| Fruits2 | Dafeng | 1 | 1.02 (0.42-2.47) | 0.42 (0.16-1.12) | 1.23 (0.51-2. 98) | 0.802 |
| Ganyu | 1 | 1.61 (0.68-3.80) | 1.13 (0.43-2.95) | 1.17 (0.41-3.37) | 0.746 | |
Matched by age and gender, further adjusted for level of education, past economic status (group), smoking, alcohol drinking, BMI group, cancer family history, eating speed, food temperature and self-reported grain fungi pollution;
Matched by age and gender, further adjusted for education level, past economic status (group), smoking, alcohol drinking, BMI group, eating speed and family history of cancer;
Less than 3 times per week=1, 3 times per week and above=2;
Matched by age and gender, further adjusted for level of education, past economic status (group), smoking, alcohol drinking, BMI group, cancer family history, eating speed and food temperature;
Categorized by median among controls.
With regard to the consumption of major food groups, high consumptions of fish and seafood products significantly elevated the risk of developing EC in Dafeng (for trend P = 0.024). Staple foods, preserved foods, fruits and vegetables, and soybean, however, were not apparently associated with EC risk in either county.
DISCUSSION
This population-based case-control study, conducted in high and low-risk areas of Jiangsu Province, China demonstrated an association between tobacco smoking, alcohol drinking, dietary factors and EC. These associations were compared separately in the two regions which had similar socioeconomic status and geographic characteristics. From our awareness, this is the first comparative population-based case-control study conducted in low-risk and high-risk areas simultaneously to compare the different associations of risk factors and EC in similar areas with high-risk gradients. Like other epidemiological researches, our study showed that EC was inversely associated with socioeconomic status, such as level of education and income. However this disease was more frequent among subjects with smoking and alcohol drinking habits and unhealthy dietary factors. Furthermore, a geographic variation of some associations was observed between the low-risk and high-risk areas. Smoking elevated the risk of EC in both areas concordantly, but the dose-response relationship of smoking and smoking related variables (age of first smoking, duration and amount) was apparent only in the high-risk area. On the contrary, the effect of alcohol drinking on EC and a dose-response relationship was only observed in the low-risk area.
Supporting previous studies, the risk of EC was inversely associated with socioeconomic status in the present study[15,16]. People with higher levels of education and better financial situations tend to have a lower risk of developing EC due to good living conditions and better health care access. Increased risk was found in people who had encountered misfortune in the past 10 years (Dafeng OR: 1.26, Ganyu OR: 1.64), or had a history of family cancer in first-degree relatives (Dafeng OR: 1.53, Ganyu OR: 1.57). These results were consistent in both high-risk and low-risk areas.
Associations between body mass index (BMI) and EC have been explored in several studies. Chow
et al[17] reported a tendency towards a decreasing risk of esophageal squamous cell carcinoma with increasing BMI. Engeland found that low BMI increased the risk of esophageal squamous cell carcinoma, while high BMI increased the risk of esophageal adenocarcinoma. In general, lowest BMI had the highest risk of EC[18]. Our study found similar results in both low-risk and high-risk areas. The risk of developing EC was significantly lower in normal and overweight groups when compared to the underweight group. However, the OR was high in the obese group. An increased risk of esophageal adenocarcinoma among obese persons has been explained by a dose-dependent association between increasing BMI and the risk of gastro-esophageal reflux symptoms, as observed by Nilsson et al[19].
In conformity with other epidemiological studies shown in Western countries and some areas of Asia and Africa[20-22], increased risks of EC among former smoking and current smoking subjects were observed in both areas of our study. Tobacco smoke contains over 3000 constituents including 30 carcinogens, such as polycyclic aromatic hydrocarbons (PAHs), aromatic amines, and N-nitrosamines. The metabolites of these carcinogens may lead to gene mutation and cancer[23]. Age of first smoke, duration and dosage of tobacco use were also strongly associated with an elevated risk of developing EC in Dafeng, with an apparent dose-response relationship. Although Ganyu had a similar smoking prevalence, these time and dosage dependent results were not statistically significant in this area.
Several studies have reported a strong correlation between EC and alcohol abuse[24,25]. Alcoholic beverages also contain carcinogens and other compounds and may facilitate the absorption of esophageal mucosal cells and make them more susceptible to chemical carcinogens[26]. On the contrary of tobacco smoking, the positive association between alcohol drinking and EC was only found in Ganyu in our study (OR=1.71). Several studies have reported a linear relationship between an overall daily ethanol consumption and EC risk[27,28]. However, in our study only the age of initial drinking and years of alcohol drinking were found to be associated with EC risk in Ganyu. No clear relationship between daily alcohol consumption and EC was found.
The interaction between tobacco smoking and alcohol drinking has been studied in many researches. It has been suggested that alcohol and tobacco interact in a multiplicative way[29,30]. In a large scale study, Castellsagué reported that the risk of EC in the highest joint level of alcohol and cigarette smoking increased 50.85-fold and 35.34-fold among men and women[7]. However the joint effect of smoking and alcohol was not found to be statistically significant in the high-risk area and the low-risk area in our study. The link between smoking, alcohol and EC in China are not as apparent as in Western countries. Several previous studies conducted in other high-risk areas of Jiangsu, China either did not find any relation or found only a weak association between smoking, alcohol drinking and EC[31,32].
Dietary factors are thought to play an important role in the pathogenesis of EC. Some epidemiological studies have suggested that the risk of EC is inversely associated with a higher intake of fruits and vegetables[10,33], while a detrimental effect was observed among high intake of certain types of meat, butter and saturated fatty acids[34]. Increased risk was related to N-nitrosamine compounds (mainly from preserved foods), foods contamination by fungus and the presence of toxins. Some unhealthy dietary habits such as fast eating speeds, consumption of hot foods and soups can cause the injury of esophageal mucosa and render the mucosa more susceptible to carcinogens.
An increased OR was found among fast eating subjects in both areas of this study (Dafeng OR: 4.01, Ganyu OR: 3.09). However, the associations between high food temperatures, the possibility of fungi pollution of grain, frequent intake of fresh garlic and EC was not statistically significant. In the food group analysis, after adjusting for potential confounders, we did not find any significant association between major food consumption and EC risk in either area. A positive association of fish and seafood product intake in Dafeng was found (for trend P = 0.024). Fish is a rich dietary source of n-3 fatty acids. It has been reported that this long chain of fatty acid can suppress mutation, inhibit cell growth, and enhance cell apoptosis, thus reducing the risk of developing cancer[35]. The contradicting results found in our study of the increased risk found in the association between fish consumption and EC in Dafeng was probably due to water contamination or other unidentified confounders. However, this hypothesis needs to be further clarified and studied. Moreover, it may be more reasonable to study food composition and micronutrients in our future analysis rather than to use individual foods or food groups[36].
Ganyu has a high proportion of ageing and illiterate residents. The economic status of residents in Ganyu is also lower than Dafeng, although the difference is not significant (Table 1). As disease is more prevalent among ageing populations and the level of education and economic situation are inversely associated with the risk of developing EC, it can be expected that the two counties would have a far higher risk gradient if they had a similar distribution of age and socioeconomic related factors.
As mentioned above, a heterogeneous association between smoking, alcohol drinking and EC was observed in the low- and high-risk areas in our study, despite their similar geographic characteristics and general socioeconomic statuses. A malignant tumor is the result of a series of DNA alterations in a single cell, which leads to a loss of normal functioning. A large number of gene coding for enzymes and receptors are involved in xenobiotic metabolism, with many of them showing polymorphisms. Many molecular epidemiological studies have proved that polymorphisms in activation and detoxification enzymes can interact with environmental carcinogens. It has been reported that GSTM1 null carriers may be especially susceptible to the action of tobacco with regards to EC[32], while inactive the ALDH2 genotype increases the risk of EC in alcoholics[37]. Genetic polymorphisms can interact with dietary factors. For example, cruciferous vegetables can inhibit the metabolic activation of phase I enzymes and induce the detoxification of carcinogens via phase II enzymes[38]. The polymorphism of one gene may also have an effect on other genes. Gene-gene interactions between GSTM1 0/0 and CYP1A1 and CYP1A2 enzyme induction have been observed in smokers[39]. Another example is that individuals with CYP1A1 Ile/Val alleles have greater CYP1A2 activity than those with wild type CYP1A1[40]. Furthermore, it has been suggested that genes can influence individual behaviours such as smoking, alcohol drinking and excess calorie intake, thereby having the potential to affect cancer risk[41].
Both environmental factors and human genes can show considerable regional variability. The variation in these factors together with their separate and joint effects ultimately determine the risk of cancer in different regions and may be the main reason for the large EC risk gradient between the counties in Jiangsu Province. Unfortunately scientific evidence on genetic polymorphisms, gene-environmental and gene-gene interactions remains inconsistent and inconclusive because of low statistical power and few candidate genes in previous studies. Moreover, no study has ever been conducted to compare the association between gene-environmental interaction and EC risk in apparently similar areas with a high risk gradient. Therefore, our future study will focus on genetic polymorphisms and their interactions with different environmental, lifestyle and dietary factors in the etiology of EC in high and low-risk areas, with a sufficient sample size and candidate genes.
Our present population-based case-control study has some limitations. Differences in the etiological factors between esophageal adenocarcinoma and squamous cell carcinoma may exist. Because of the low histological examination rate in China, it is difficult to differentiate between the subtypes of EC in a population-based study. Additionally, most risk factors in our study are based on self-reported data and may be subject to recall bias. Moreover, the relationship between BMI and EC was examined by using height and weight measurements obtained at the time of interview. Some cases might have begun to lose weight at an earlier time because of the disease. This factor could also have caused bias in our study.
In summary, the present study demonstrated the association between smoking, alcohol drinking, dietary factors and EC risk in the low-risk and high-risk areas of Jiangsu Province, China. Heterogeneous effects of smoking and alcohol drinking were found between the two areas, despite their similar geographic characteristics and general socioeconomic status. The variation in environmental risk factors, together with gene-environment and gene-gene interactions may be the main reason for these heterogeneous associations and may contribute to the large risk gradient of EC mortality in Jiangsu Province, China.
ACKNOWLEDGMENTS
The staff of local Health Bureau and local CDC in Dafeng and Ganyu County for their assistance in fieldworks.
Footnotes
Supported by Jiangsu Provincial Health Department No. RC 2003090
S- Editor Pan BR L- Editor Zhang JZ E- Editor Ma WH
References
- 1.Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin. 2005;55:74–108. doi: 10.3322/canjclin.55.2.74. [DOI] [PubMed] [Google Scholar]
- 2.Ke L. Mortality and incidence trends from esophagus cancer in selected geographic areas of China circa 1970-90. Int J Cancer. 2002;102:271–274. doi: 10.1002/ijc.10706. [DOI] [PubMed] [Google Scholar]
- 3.Pu ST, Xiong ZF. Total death analysis in Jiangsu from 1990 to 1992. Nanjing: South-east University; 1996. [Google Scholar]
- 4.Hu XS, Zhou XN, Sun NS, Zhao JK, Wu M, Wang PH, Yang GJ. Spatial analysis of distribution of malignant neoplasm using GIS in Jiangsu, China. Zhonghua Liuxing Bingxue Zazhi. 2002;23(1):73–4. [Google Scholar]
- 5.Tobacco smoking. IARC Monogr Eval Carcinog Risk Chem Hum. 1986;38:35–394. [PubMed] [Google Scholar]
- 6.Alcohol drinking. Biological data relevant to the evaluation of carcinogenic risk to humans. IARC Monogr Eval Carcinog Risks Hum. 1988;44:101–152. [PMC free article] [PubMed] [Google Scholar]
- 7.Castellsague X, Munoz N, De Stefani E, Victora CG, Castelletto R, Rolon PA, Quintana MJ. Independent and joint effects of tobacco smoking and alcohol drinking on the risk of esophageal cancer in men and women. Int J Cancer. 1999;82:657–664. doi: 10.1002/(sici)1097-0215(19990827)82:5<657::aid-ijc7>3.0.co;2-c. [DOI] [PubMed] [Google Scholar]
- 8.Chainani-Wu N. Diet and oral, pharyngeal, and esophageal cancer. Nutr Cancer. 2002;44:104–126. doi: 10.1207/S15327914NC4402_01. [DOI] [PubMed] [Google Scholar]
- 9.Hu J, Nyrén O, Wolk A, Bergström R, Yuen J, Adami HO, Guo L, Li H, Huang G, Xu X. Risk factors for oesophageal cancer in northeast China. Int J Cancer. 1994;57:38–46. doi: 10.1002/ijc.2910570108. [DOI] [PubMed] [Google Scholar]
- 10.Phukan RK, Chetia CK, Ali MS, Mahanta J. Role of dietary habits in the development of esophageal cancer in Assam, the north-eastern region of India. Nutr Cancer. 2001;39:204–209. doi: 10.1207/S15327914nc392_7. [DOI] [PubMed] [Google Scholar]
- 11.Dong Z, Tang P, Li L, Wang G. The strategy for esophageal cancer control in high-risk areas of China. Jpn J Clin Oncol. 2002;32 Suppl:S10–S12. doi: 10.1093/jjco/hye122. [DOI] [PubMed] [Google Scholar]
- 12.Zhao JK, Liu AM, Wang XS, Wu M, Sheng LG, Lu J. An Analysis on Death Cause of Cancer in High and Low Incidence Areas of Jiangsu Province. Zhongguo Zhongliu. 2004;13(12):757–9. [Google Scholar]
- 13.Zhou B. Predictive values of body mass index and waist circumference to risk factors of related diseases in Chinese adult population. Zhonghua Liuxing Bingxue Zazhi. 2002;23:5–10. [PubMed] [Google Scholar]
- 14.Breslow NE, Day NE. Statistical methods in cancer research. Volume I - The analysis of case-control studies. IARC Sci Publ. 1980;32:5–338. [PubMed] [Google Scholar]
- 15.Brown LM, Silverman DT, Pottern LM, Schoenberg JB, Greenberg RS, Swanson GM, Liff JM, Schwartz AG, Hayes RB, Blot WJ. Adenocarcinoma of the esophagus and esophagogastric junction in white men in the United States: alcohol, tobacco, and socioeconomic factors. Cancer Causes Control. 1994;5:333–340. doi: 10.1007/BF01804984. [DOI] [PubMed] [Google Scholar]
- 16.Ward MH, Dosemeci M, Cocco P. Mortality from gastric cardia and lower esophagus cancer and occupation. J Occup Med. 1994;36:1222–1227. doi: 10.1097/00043764-199411000-00008. [DOI] [PubMed] [Google Scholar]
- 17.Chow WH, Blot WJ, Vaughan TL, Risch HA, Gammon MD, Stanford JL, Dubrow R, Schoenberg JB, Mayne ST, Farrow DC, et al. Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst. 1998;90:150–155. doi: 10.1093/jnci/90.2.150. [DOI] [PubMed] [Google Scholar]
- 18.Engeland A, Tretli S, Bjørge T. Height and body mass index in relation to esophageal cancer; 23-year follow-up of two million Norwegian men and women. Cancer Causes Control. 2004;15:837–843. doi: 10.1023/B:CACO.0000043434.21558.ea. [DOI] [PubMed] [Google Scholar]
- 19.Nilsson M, Johnsen R, Ye W, Hveem K, Lagergren J. Obesity and estrogen as risk factors for gastroesophageal reflux symptoms. JAMA. 2003;290:66–72. doi: 10.1001/jama.290.1.66. [DOI] [PubMed] [Google Scholar]
- 20.Yang CX, Wang HY, Wang ZM, Du HZ, Tao DM, Mu XY, Chen HG, Lei Y, Matsuo K, Tajima K. Risk factors for esophageal cancer: a case-control study in South-western China. Asian Pac J Cancer Prev. 2005;6:48–53. [PubMed] [Google Scholar]
- 21.Gallus S, Altieri A, Bosetti C, Franceschi S, Levi F, Negri E, Dal Maso L, Conti E, Zambon P, La Vecchia C. Cigarette tar yield and risk of upper digestive tract cancers: case-control studies from Italy and Switzerland. Ann Oncol. 2003;14:209–213. doi: 10.1093/annonc/mdg074. [DOI] [PubMed] [Google Scholar]
- 22.Wu AH, Wan P, Bernstein L. A multiethnic population-based study of smoking, alcohol and body size and risk of adenocarcinomas of the stomach and esophagus (United States) Cancer Causes Control. 2001;12:721–732. doi: 10.1023/a:1011290704728. [DOI] [PubMed] [Google Scholar]
- 23.Hecht SS. Tobacco smoke carcinogens and lung cancer. J Natl Cancer Inst. 1999;91:1194–1210. doi: 10.1093/jnci/91.14.1194. [DOI] [PubMed] [Google Scholar]
- 24.Yokoyama A, Kato H, Yokoyama T, Tsujinaka T, Muto M, Omori T, Haneda T, Kumagai Y, Igaki H, Yokoyama M, et al. Genetic polymorphisms of alcohol and aldehyde dehydrogenases and glutathione S-transferase M1 and drinking, smoking, and diet in Japanese men with esophageal squamous cell carcinoma. Carcinogenesis. 2002;23:1851–1859. doi: 10.1093/carcin/23.11.1851. [DOI] [PubMed] [Google Scholar]
- 25.Chao YC, Wang LS, Hsieh TY, Chu CW, Chang FY, Chu HC. Chinese alcoholic patients with esophageal cancer are genetically different from alcoholics with acute pancreatitis and liver cirrhosis. Am J Gastroenterol. 2000;95:2958–2964. doi: 10.1111/j.1572-0241.2000.02328.x. [DOI] [PubMed] [Google Scholar]
- 26.Mufti SI, Becker G, Sipes IG. Effect of chronic dietary ethanol consumption on the initiation and promotion of chemically-induced esophageal carcinogenesis in experimental rats. Carcinogenesis. 1989;10:303–309. doi: 10.1093/carcin/10.2.303. [DOI] [PubMed] [Google Scholar]
- 27.Lee CH, Lee JM, Wu DC, Hsu HK, Kao EL, Huang HL, Wang TN, Huang MC, Wu MT. Independent and combined effects of alcohol intake, tobacco smoking and betel quid chewing on the risk of esophageal cancer in Taiwan. Int J Cancer. 2005;113:475–482. doi: 10.1002/ijc.20619. [DOI] [PubMed] [Google Scholar]
- 28.Castelletto R, Castellsague X, Muñoz N, Iscovich J, Chopita N, Jmelnitsky A. Alcohol, tobacco, diet, mate drinking, and esophageal cancer in Argentina. Cancer Epidemiol Biomarkers Prev. 1994;3:557–564. [PubMed] [Google Scholar]
- 29.De Stefani E, Muñoz N, Estève J, Vasallo A, Victora CG, Teuchmann S. Mate drinking, alcohol, tobacco, diet, and esophageal cancer in Uruguay. Cancer Res. 1990;50:426–431. [PubMed] [Google Scholar]
- 30.Zeka A, Gore R, Kriebel D. Effects of alcohol and tobacco on aerodigestive cancer risks: a meta-regression analysis. Cancer Causes Control. 2003;14:897–906. doi: 10.1023/b:caco.0000003854.34221.a8. [DOI] [PubMed] [Google Scholar]
- 31.Takezaki T, Gao CM, Wu JZ, Ding JH, Liu YT, Zhang Y, Li SP, Su P, Liu TK, Tajima K. Dietary protective and risk factors for esophageal and stomach cancers in a low-epidemic area for stomach cancer in Jiangsu Province, China: comparison with those in a high-epidemic area. Jpn J Cancer Res. 2001;92:1157–1165. doi: 10.1111/j.1349-7006.2001.tb02135.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gao CM, Takezaki T, Wu JZ, Li ZY, Liu YT, Li SP, Ding JH, Su P, Hu X, Xu TL, et al. Glutathione-S-transferases M1 (GSTM1) and GSTT1 genotype, smoking, consumption of alcohol and tea and risk of esophageal and stomach cancers: a case-control study of a high-incidence area in Jiangsu Province, China. Cancer Lett. 2002;188:95–102. doi: 10.1016/s0304-3835(02)00115-5. [DOI] [PubMed] [Google Scholar]
- 33.Norat T, Riboli E. Fruit and vegetable consumption and risk of cancer of the digestive tract: meta-analysis of published case-control and cohort studies. IARC Sci Publ. 2002;156:123–125. [PubMed] [Google Scholar]
- 34.De Stefani E, Deneo-Pellegrini H, Boffetta P, Mendilaharsu M. Meat intake and risk of squamous cell esophageal cancer: a case-control study in Uruguay. Int J Cancer. 1999;82:33–37. doi: 10.1002/(sici)1097-0215(19990702)82:1<33::aid-ijc7>3.0.co;2-7. [DOI] [PubMed] [Google Scholar]
- 35.Du C, Fujii Y, Ito M, Harada M, Moriyama E, Shimada R, Ikemoto A, Okuyama H. Dietary polyunsaturated fatty acids suppress acute hepatitis, alter gene expression and prolong survival of female Long-Evans Cinnamon rats, a model of Wilson disease. J Nutr Biochem. 2004;15:273–280. doi: 10.1016/j.jnutbio.2003.11.005. [DOI] [PubMed] [Google Scholar]
- 36.Vastag B. Recent studies show limited association of fruit and vegetable consumption and cancer risk. J Natl Cancer Inst. 2005;97:474–476. doi: 10.1093/jnci/97.7.474. [DOI] [PubMed] [Google Scholar]
- 37.Matsuo K, Hamajima N, Shinoda M, Hatooka S, Inoue M, Takezaki T, Tajima K. Gene-environment interaction between an aldehyde dehydrogenase-2 (ALDH2) polymorphism and alcohol consumption for the risk of esophageal cancer. Carcinogenesis. 2001;22:913–916. doi: 10.1093/carcin/22.6.913. [DOI] [PubMed] [Google Scholar]
- 38.Murillo G, Mehta RG. Cruciferous vegetables and cancer prevention. Nutr Cancer. 2001;41:17–28. doi: 10.1080/01635581.2001.9680607. [DOI] [PubMed] [Google Scholar]
- 39.MacLeod S, Sinha R, Kadlubar FF, Lang NP. Polymorphisms of CYP1A1 and GSTM1 influence the in vivo function of CYP1A2. Mutat Res. 1997;376:135–142. doi: 10.1016/s0027-5107(97)00036-5. [DOI] [PubMed] [Google Scholar]
- 40.Bartsch H. DNA adducts in human carcinogenesis: etiological relevance and structure-activity relationship. Mutat Res. 1996;340:67–79. doi: 10.1016/s0165-1110(96)90040-8. [DOI] [PubMed] [Google Scholar]
- 41.Freedman AN. Somatic alterations and metabolic polymorphisms. IARC Sci Publ. 1999;148:37–50. [PubMed] [Google Scholar]