The Relative Factors Of Environment And Diet Involved In Esophageal Cancer In Eastern Guangdong

BackgroundEsophageal cancer (EC) is a common malignant tumor in human digestive tract. China is a country with the highest mortality rate of EC. The highest EC mortality in eastern Guangdong is more than 5 times the Chinese average. Therefore, the integrative role of specific environmental factors and lifestyle of local inhabitants on the EC etiology must be further evaluated. On the EC epidemiology, previous studies had failed to use the methods of data mining method and field investigation to explore the correlation among lifestyle, environmental condition, socioeconomic status and EC geographical distribution.In this report, we used the methods of data mining method and field investigation to explore the correlation among environmental-social factors, nitrogen pollution levels, consumption of tobacco-alcohol, intake of salted food and dietary pattern in the areas of different EC risk. By multiple linear and nonlinear models, the relationship among the character of the geographic distribution of mortality rates of EC and the integrative effects of multiple factors involved in the specific environmental factors and lifestyle in eastern Guangdong was studied and the hypothesis of environmental factors regarding EC causes was also verified.MethodsStudy subjects were recruited from 9 counties/cities, i.e. Nan’ao, Raoping, Jieyang, Chaozhou, Chenghai, urban of Shantou, Chaoyang, Lufeng and Haifeng, in eastern Guangdong. The EC mortality rates, specific environmental factors, socioeconomic status and lifestyle of the inhabitants in above counties/cities were surveyed. The standardized mortality rates (SMR) of EC of the counties/cities, adjusted by China standard population in 1964, were provided by China Cancer Database in 1970’s (1973-1975). Environmental and social economic conditions i.e.①E nvironmental pollution condition,②climate geographical factors,③land and water resources situation,④t he social economic conditions, were obtained by data mining method. Then, five thousands of the residents were randomly chosen from above areas to investigate their living and dietary habits. The main indexes of survey included Smoking Index (SI), Alcohol -drinking Index (AI), Body Mass Index (BMI), etc. Based on the characteristics of different data, various modeling methods were performed to construct the multi-factor model regarding hypothesis of environmental causes of EC. Main models include Ordinal Regression Logit Model (PLUM), Structured Multiphase Regression Model, Factor Analysis-Logarithmic Regression Model and BP Neural Network Model (BPNN). Various analytical approaches were performed to identify the comprehensive factors of the environment and lifestyle involved in the EC mortality in Chaoshan area, and to evaluate the potential effects. SPSS15.0 and SPSS clementine10.0 were used to perform statistic analyses.Results1. Lifestyle of the inhabitants from different areas in eastern GuangdongThe questionnaire is reliable with both of Kappa and Spearman correlation coefficient >0.7. In eastern Guangdong, the smoking ratio (66.5%) of male inhabitants was more 66.0% than that of female (0.5%) while male drinking ratio (84.0%) was more 83.7% than that (0.3%). Each of the smoking ratio and drinking ratio in Raoping county was the highest in 9 counties/cities (P<0.05). Pickle intake frequency (35.9±1.28 = mean±s, day/year) in Jieyang county was the highest in 9 counties/cities (P<0.05). On the other hand, the inhabitants in Lufeng had the highest intake frequency of salted fish (35.3±1.5, P<0.05).The mode and concentration of drinking tea in eastern Guangdong were different with highly statistical significance (both P=0.00).2. Identification of relative Factors (principal component) involved in the EC mortality in eastern GuangdongThe results of Factor Regression Model, Factor Analysis-Logarithmic Regression Model and Factor Analysis-Ordinal Regression Model showed that the most important risk factor involved in male or female EC death in eastern Guangdong in 1970’s was Environment-Pollution Factor. Environment-Pollution Factor was mainly composed of Nitrogen-load Index regarding pig waste,the consumption of pesticide per hectare farmland, and nitrogenous fertilizer. Factor Analysis-Logarithmic Regression Model showed that male and female EC-SMR might increase 1.92/105 and 2.06/105 respectively when Environment-Pollution Factor increased 1 unit.3. Identification of relative factors involved in the EC mortality in eastern Guangdong⑴Structured Multiphase Regression Model The potential risk factors of male EC death in eastern Guangdong were average altitude, Nitrogen-load Index regarding pig waste, tea drinking mode (1= No-kungfu tea, 2=kungfu tea with No-kungfu tea, 3=kungfu tea), the total consumption of cooking animal oil, the cumulative consumption of tobacco; The potential protective factor was the consumed frequency of seafood. Y(SMR-M) = -4.493 +2.693×altitude +0.442×Nitrogen-load Index -0.604×consumed frequency of seafood+8.122×tea drinking mode +0.230×total consumption of cooking animal oil +0.001×cumulative consumption of tobacco (p=0.000, AdjR~2=0.640 for model). Nitrogen-load Index might be a most important factor among 6 factors affected Y (SMR-M) and explain 46.4% of variance for the male EC-SMR. In addition, the potential risk factors involved in female EC death in eastern Guangdong were average altitude, Nitrogen-load Index, tea drinking mode, the total consumption of cooking oil and pickle consumption. The protective factor was the consumed frequency of seafood. Y(SMR-F)=-5.998 +1.182×altitude +0.221×Nitrogen-load Index -0.303×consumed frequency of seafood +4.086×tea drinking mode +0.095×total consumption of cooking oil +0.016×pickle consumption (P=0.000,AdjR~2=0.715 for model). Nitrogen-load Index might be a most important factor among 6 factors affected Y (SMR-F) and explain 56.4% of variance for the female EC-SMR.⑵Factor Analysis-Logarithmic Regress Model The potential risk factors involved in male EC death in eastern Guangdong were average altitude (OR=1.05), Nitrogen-load Index (OR=1.01), tea drinking mode (OR=1.51), whether drinking frequently (OR=3.39), the total consumption of cooking animal oil (OR=1.01), the cumulative consumption of alcohol (OR=1.01), daily consumption of tobacco (OR=1.25) and pickle consumption (OR=1.01). The protective factor was the consumed frequency of seafood (OR=0.98). The AdjR 2 for model was 0.516. Also, the potential risk factors of female EC death in eastern Guangdong were average altitude (OR=1.03), Nitrogen-load Index (OR=1.01), tea drinking mode (OR=1.44), the total consumption of cooking oil (OR=1.01), pickle consumption (OR=1.01), consumption mode of fish sauce (OR=1.44) and tea temperature (OR =1.01). The potential protective factor was also the consumed frequency of seafood (OR =0.97). The Adj R~2 for the model was 0.462. Each of male and female EC-SMR might increase 1.01/105 when Nitrogen-load Index regarding pig waste increased 1 unit.⑶Ordinal Regression Logit Model In eastern Guangdong, the potential risk factors involved in male EC death were average altitude (OR=1.16), Nitrogen-load Index (OR=1.04), the cumulative consumption of alcohol (OR=1.01), the total consumption of cooking animal oil (OR=1.02), tea drinking mode (0.37 for kungfu tea with No-kungfu tea and 0.25 for No-kungfu tea, reference level=kungfu tea) and no-smoking (OR=0.75). The protective factor was the consumed frequency of seafood (OR=0.96). In female EC death in eastern Guangdong, the potential risk factors involved were average altitude (OR=1.33), Nitrogen-load Index (OR=1.05), the total consumption of cooking oil (OR=1.02) and tea drinking mode (0.29 for kungfu tea with No-kungfu tea and 0.21 for No-kungfu tea, reference level =kungfu tea). The protective factor was also the consumed frequency of seafood (OR=0.93).⑷BPNN Model On top, the three factors involved in EC-SMR were also the consumption of pesticide and nitrogenous fertilizer per hectare farmland, and Nitrogen-load Index regarding pig waste. The prediction accuracy of EC-SMR by BPNN was 97.82%.ConclusionsIn the natural environment, Nitrogen-load Index regarding pig waste,the consumption of pesticide and nitrogenous fertilizer per hectare farmland, and average altitude might be risk factors of EC death in eastern Guangdong in 1970’s.As to the dietary pattern, tobacco and alcohol consumption, pickle consumption, tea drinking mode and tea temperature might be risk factors of EC death in eastern Guangdong in 1970’s. The smoking and drinking might play an important role in the male EC death. Seafood may be a protective factor involved in EC in local residents.