Author Correction: Investigation of Dietary Factors and Esophageal Cancer Knowledge: Comparison of Rural Residents in High- and Low-incidence Areas

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Statistical Analysis. Statistical analyses were performed using SPSS 22.0 software (SPSS, Inc., Chicago, IL, USA). The data are reported as the frequencies, means, and medians with percentages. A chi-square test was used to compare the categorical variables. For all statistical tests, a P-value < 0.05 was considered statistically significant.  Table 1. Personal dietary habits of the rural residents in the high-and low-EC incidence areas. The survey defines the tastes of the respondents by referring to the family members' evaluations. The degree of saltiness was depended on the median sodium intake, that was slightly salty <1 g/day, moderately salty 1-3 g/day and very salty >3 g/day, respectively. The "eating speed" was described as "rapid", which indicated finishing their meal in 10 minutes; "usual", which indicated finishing their meal in "30 minutes"; or "slow", which indicated finishing their meal in "1 hour or more". The "food temperature" was described as "hot" (over 60 °C), "cold" (less than 5 °C), or "moderate temperature" (5 °C-60 °C). In terms of the family dietary status of the rural residents in the high-and low-EC incidence areas in the past year, the residents in the high-incidence area usually ate fresh foods, such as vegetables/fruits and beans, accounting for 37.3% and 31.0% of the participants, respectively, and the corresponding proportions of participants in the low-incidence area were 37.4% and 25.2%, respectively, in the past year. There were no significant differences between the two groups (P > 0.05). In the past year, the proportions of pickled vegetables, pickled meat, and barbecued food that were consumed by the high-and low-incidence area residents were significantly different (P < 0.05). Residents in Yanting consumed more pickled vegetables, salted meat and barbecued food but similar amounts of fresh vegetables/fruits, beans, sauerkraut, hot food, and spicy food compared with those of Qingzhen County in the past year (Table 2).
A comparison of the high-and low-incidence areas suggests that the proportions of residents who gradually increased their consumption of vegetables/fruits, beans, pickled vegetables, pickled meat, sauerkraut, hot food, spicy food and barbecued food were 26 Table 2. Family dietary status of the rural residents in the high-and low-EC incidence areas in the past year. We designed the degree of spicy food with the Scoville rating of 300,000. That meant participants who intake the habanero peppers. A check signifies having eaten the food every day and is defined by the term "many", an average intake per month of no more than 1 time is denoted as "few", and other values are defined as "moderate". *P < 0.05 is statistically significant.
SCientifiC REpoRTS | (2018) 8:4914 | DOI:10.1038/s41598-018-23251-3 significant differences between the two groups (P < 0.05). The past ten-year trend in Yanting revealed that the proportions of residents who consumed fresh vegetables/fruits, beans, sauerkraut, hot food, and barbecued food with a gradually increasing trend were less than those of residents in Qingzhen County. However, the gradually increasing trends in consumption of pickled vegetables, pickled meat, and spicy food in Yanting County in the past 10 years were more than those in Qingzhen County (Table 3). In the high-EC incidence area, the main source of drinking water was tap water which in the permissible value of Drinking Water Quality Standards in China, accounting for 72.5% of the residents. In the low-incidence area, the main source of drinking water was also tap water, accounting for 59.1% of the residents. Compared with the drinking water of the residents in the high-EC incidence area, that of the residents in the low-incidence area used less tap water. The drinking water sources of the two areas were significantly different (P < 0.05) ( Table 4).
The EC knowledge among rural residents in the high-and low-incidence areas was also surveyed in our study. The percentages of residents in the high-vs. low-incidence areas who had heard of EC and those with a clear understanding of the common causes, main symptoms, therapeutic measures, preventive measures, and government interventions were 17.0% vs. 4.5%, 12.1% vs. 4.5%, 13.5% vs. 3.8%, 7.8% vs. 1.9%, 8.9 vs. 1.9%, and 6.9% vs. 1.9%, respectively. All proportions were higher in the high-incidence are than in the low-incidence area. There were statistically significant differences between the two areas (P < 0.05) ( Table 5).

Discussion
EC is the result of a combination of causes and is not associated with only a single factor 17 . Because of the coordination effect between EC and factors such as the living environment, dietary habits, smoking history, alcohol drinking, low social and economic status, poor standard of living and inadequate knowledge related to EC 18 Table 3. Family dietary status of the rural residents in the high-and low-EC incidence areas in the past 10 years. Δ Each questionnaire surveyed the food intake of only one household over 10 years. *P < 0.05 is statistically significant. Many epidemiological studies have shown there is an important relationship between EC occurrence and dietary status 7,19-21 . Residents with EC in a high-incidence area have diets characterized by hot food, dry food, pickled vegetables, barbecued food, and spicy food, which can cause chronic physical and chemical damage to the esophageal mucosa, increasing its susceptibility to carcinogenic factors and carcinogenesis 13,22 . In our study, the proportions of participants eating salty food, hard food, eating rapidly, and hot food in the high-incidence area were higher than those in the low-incidence are. However, there was no significant difference in personal eating habits between the rural residents in the two areas (P > 0.05). This finding may be due to the dietary habits of residents in both areas.
Pickled vegetables, pickles, sauerkraut and barbecued food contain N-nitroso-compounds (NOCs), which have been found in animal models to induce malignant tumors, including EC 23,24 . Epidemiological studies have also shown that the NOC content in food in high-incidence areas is remarkably higher than that in the low-incidence areas of China 23 . There is a significant positive correlation between the NOC level in food and EC morbidity 25 . NOCs may play an important role in EC carcinogenesis in China. In terms of family diet in the past year, the proportions of residents who consumed high amounts of pickled vegetables, pickles, and barbecued food in the high-incidence area were significantly higher than those in the low-incidence area (P < 0.05). Additionally, in the past 10 years, the consumption of pickled vegetables and pickled meat, which contain NOCs, was also significantly higher in the high-incidence area than in the low-incidence area (P < 0.05). Eating more vegetables/ fruits and beans is a protective factor against EC because they increase the intake of anti-carcinogenic substances and reduce the incidence of EC 7,20,21 . In our study, there were no significant differences in the consumption of vegetables/fruits and beans between the two areas in the past year (P > 0.05). However, in the past 10 years, the proportion of vegetable/fruit intake has gradually increased, and the consumption of beans by residents was significantly lower in the high-incidence area than in the low-incidence area (P < 0.05). In the past year and in the past 10 years, the high-incidence area residents tended to consume more harmful foods. Our results showed that although many protective measures were performed in Yanting, changing the dietary habits of the residents in the area is difficult, which might explain why Yanting remains a high-incidence area after more than twenty years of preventative measures 1 .
Many studies have demonstrated the relationship between drinking water sources and EC 19,26 . The type and quality of drinking water are related to the intake of water with "three nitrogen" content in high-EC incidence areas [26][27][28] . Well water is considered to be contaminated with nitrite nitrogen that reacts with amines in vivo to form N-nitroso compounds, which are the causative factors for EC 28 . Low quality drinking water has been confirmed to be a risk factor in China in previous studies 22,29 . In our study, approximately 72.5% and 59.1% of residents in the high-and low-EC incidence areas, respectively, used tap water, which contains less contaminants, as the main source of drinking water. The drinking water sources in the two areas were significantly different (P < 0.05). The water quality in the high-incidence area was demonstrably better than that in the low-incidence area. This finding is related to government measures to implement projects that improved tap water quality, which were considered to be protective factor and had been popular in Yanting County.
Many previous studies have shown a 5-year survival rate of more than 95% for patients with early-stage EC [30][31][32] . However, EC lacks typical symptoms at the early stage of the disease, and it can take 2-3 years from the time of onset to the manifestation of symptoms. Once admitted to the hospital, most patients are in the advanced stage of disease and have thus missed the best time for treatment 32 . Prevention and treatment of EC in the high-incidence area of Yanting has been performed for many years 11 . However, knowledge of the prevention and control of EC was still unclear in high-incidence areas. No study concerning the differences in EC knowledge between highand low-incidence areas has been conducted previously. In our study, we surveyed the knowledge of EC in both high-and low-incidence areas. Our survey demonstrated that there were statistically significant differences in whether residents had ever heard about EC and their understanding of the common causes, main symptoms, therapeutic measures, preventive measures, and government intervention in the high-versus low-incidence areas (P < 0.05). Residents in the high-incidence area had better knowledge of EC than those in the low-incidence area in these aspects. This finding may be related to propaganda from the government and disease propagandists who popularized the knowledge of EC in high-incidence areas. Besides, it may because participants from high  Table 4. Family drinking water sources of the rural residents in the high-and low-EC incidence areas. Δ In the permissible value of Drinking Water Quality Standards in China. *P < 0.05 is statistically significant.
incidence area may know more knowledge of EC from other EC patients than that of low incidence area of EC. They may also know that from their family member or neighbors. To decrease the morbidity and increase the diagnostic rate of EC, early diagnosis and treatment should be performed in high-incidence areas. Popularization of EC knowledge is an important way to promote residents' awareness of preventive and curative measures in high-incidence areas 33,34 . Therefore, improving public awareness of EC should be helpful for reducing the mortality and morbidity of EC. In summary, the dietary habits and knowledge of EC in high-and low-incidence areas were different in our survey. Residents in the high-incidence area had more harmful dietary habits than those in the low-incidence area, although many preventative measures and control policies have been performed during these years in the high-incidence area. In addition, residents in the high-incidence area had greater knowledge of EC than those in the low-incidence area. However, the prevalence of knowledge regarding EC prevention and control still needs to be strengthened.  Table 5. Personal knowledge of EC among the rural residents in the high-and low-incidence areas. Δ The common causes, main symptoms, treatment measures and preventive measures for EC were found in the 8th edition of the AJCC & UICC (2017). EC, Esophageal cancer. *P < 0.05 is statistically significant.