Inverse relations between Helicobacter pylori infection and risk of esophageal precancerous lesions in drinkers and peanut consumption

Abstract

BACKGROUND

Helicobacter pylori (H. pylori) is a Gram-negative bacterium found in the upper digestive tract. Although H. pylori infection is an identified risk factor for gastric cancer, its role in esophageal squamous cell carcinoma (ESCC) remains a topic of much debate.

AIM

To evaluate the association between H. pylori infection and the risk of precancerous lesions of ESCC, and further explore the association between dietary factors and the risk of H. pylori infection.

METHODS

Two hundred patients with esophageal precancerous lesions (EPL) aged 63.01 ± 6.08 years and 200 healthy controls aged 62.85 ± 6.03 years were included in this case-control study. Epidemiological data and qualitative food frequency data were investigated. Enzyme-linked immunosorbent assay measuring serum immunoglobulin G antibodies was used to determine H. pylori seropositivity. An unconditional logistic regression model was used to assess the association between H. pylori infection and EPL risk dichotomized by gender, age, and the use of tobacco and alcohol, as well as the association between dietary factors and the risk of H. pylori infection.

RESULTS

A total of 47 (23.5%) EPL cases and 58 (29.0%) healthy controls had positive H. pylori infection. An inverse relation between H. pylori infection and the risk of EPL was found in the group of drinkers after adjustment for covariates [odds ratio (OR) = 0.32, 95% confidence interval (95%CI): 0.11-0.95]. Additionally, peanut intake was significantly associated with a decreased risk of H. pylori infection (OR = 0.39, 95%CI: 0.20-0.74).

CONCLUSION

Our study suggested that H. pylori infection may decrease the risk of EPL for drinkers in a rural adult Chinese population, and the consumption of peanut may reduce the risk of H. pylori infection. These findings should be framed as preliminary evidence, and further studies are required to address whether the mechanisms are related to the localization of lesions and alcohol consumption.

Key Words: Helicobacter pylori, Esophageal precancerous lesions, Peanut consumption, Esophageal squamous cell carcinoma

Core Tip: The association between Helicobacter pylori (H. pylori) infection and esophageal squamous cell carcinoma (ESCC) remains a topic of much debate. This study aimed to evaluate the association between H. pylori infection and the risk of precancerous lesions of ESCC, and further explore the association between dietary intake and the risk of H. pylori infection. Our findings suggested an inverse association between H. pylori infection and the risk of esophageal precancerous lesions in the group of drinkers [odds ratio (OR) = 0.32, 95% confidence interval (95%CI): 0.11-0.95]. Additionally, peanut consumption was significantly associated with a reduced risk of H. pylori infection (OR = 0.39, 95%CI: 0.20-0.74).

INTRODUCTION

Esophageal cancer and gastric cancer are upper gastrointestinal cancers that share many risk factors[1-3]. However, their associations with Helicobacter pylori (H. pylori) infection can be completely different. It has been determined that H. pylori infection is an identified risk factor for gastric cancer[4], whereas the role of H. pylori in the risk of esophageal cancer remains controversial. Previous meta-analyses summarized that H. pylori infection is likely to be related to a reduced risk of esophageal adenocarcinoma (EAC)[5-9]. One of the reliable assumptions related to this phenomenon is that H. pylori infection causes gastric atrophy and parietal cell loss, thus leading to alleviated reflux and consequently, a decreased incidence of reflux esophagitis and Barrett’s esophagus (precursor for EAC)[10-12]. However, the impact of H. pylori infection on esophageal squamous cell carcinoma (ESCC) is not well understood, and research is inconclusive as to what population may be significantly influenced[9,13-15]. Previous meta-analyses also reported that in the general population, no significant association was found between H. pylori infection and ESCC risk[6-8], whereas an inverse association was observed in the Middle East[9]. In the other populations, the inverse relationship was found to be highly associated with age, smoking status, and drinking status[15].

H. pylori is a Gram-negative bacterium found in the upper digestive tract. In spite of the fact that H. pylori infection may reduce the risk of EAC, it may also cause an adverse effect on human health. Apart from the elevated risk of gastric cancer, H. pylori infection is also etiologically related to peptic ulcers, atrophic and non-atrophic gastritis, and lymphoma associated with gastric mucosa, and is able to induce reduced bioavailability and malabsorption of nutrients including iron and vitamin B₁₂[16-18]. This case-control study aimed to investigate the association between H. pylori infection and the risk of precancerous lesions of ESCC, which is an identified early stage of carcinogenesis, and further examine the association between dietary factors and the risk of H. pylori infection.

MATERIALS AND METHODS

This study was carried out in a high-incidence area for ESCC located in Huai’an District, Huai’an City, Jiangsu Province, China, where the crude incidence rate from 1998 to 2016 was 91.85/100000[19]. As described in our previous studies[20-22], the Early Diagnosis and Early Treatment Project of Esophageal Cancer (EDETPEC) supported by the government and Cancer Foundation of China has been carried out in the endemic regions including Huai’an District since 2010. Local residents were required to undergo routine endoscopies. A detailed introduction to esophageal precancerous lesions (EPL) based on histological criteria for dysplasia and methods for EPL diagnosis has already been given in a previous study[21]. The localization of EPL was based on the definition of upper thoracic esophagus (from thoracic inlet to level of tracheal bifurcation; 18-23 cm from incisors), mid thoracic esophagus (from tracheal bifurcation midway to gastroesophageal junction; 24-32 cm from incisors), and lower thoracic esophagus (from midway between tracheal bifurcation and gastroesophageal junction to gastroesophageal junction, including abdominal esophagus; 32-40 cm from incisors)[23]. Figure 1 shows the flowchart of the study population and data collection process. This study included 200 EPL cases aged 62.85 ± 6.03 years and 200 healthy controls aged 63.01 ± 6.08 years matched by gender, age (± 2 years), and villages. The collection of epidemiological data and dietary intake data based on questionnaire method has been introduced in detail previously[21]. Subjects were required to provide the amount of beer/wine/liquor/any other alcoholic drinks consumed per day, which meant that the average alcohol units consumed per day could be estimated. Separated serum samples were obtained by centrifuging collected fasting blood samples at 3000 rpm for 5 min. Enzyme-linked immunosorbent assay (ELISA, KingMed Diagnostics Group Co., Ltd. Guangzhou, China) measuring serum immunoglobulin G (IgG) antibodies was used to determine H. pylori seropositivity. Sensitivity of the ELISA test was 97.9% [95% confidence interval (95%CI): 88.9%-99.9%] and specificity was 100% (95%CI: 86.8%-100%).

Figure 1 Flowchart of study population and sample collection.

Epidata version 3.1 (EpiData Association, Odense, Denmark) was used for inputting and validating the epidemiological data and dietary intake data. Then, SPSS version 22.0 (SPSS, Chicago, IL, United States) was used to establish a database and perform statistical analyses. Two independent samples t-test and conditional logistic regression model were used to evaluate the differences in general characteristics and potential factors between healthy controls and EPL cases, wherever appropriate. The Fisher’s exact test was used to analyze the difference in localization of EPL and H. pylori infection. An unconditional logistic regression model was used to assess the association between H. pylori infection and EPL risk dichotomized by gender, age, and tobacco and alcohol use, as well as the association between dietary factors and H. pylori infection. Covariates including gender, age, body mass index (BMI), education level, annual income, number of cigarettes per day, and alcohol units consumed per day were adjusted in the logistic regression model. Meanwhile, odds ratio (OR) and 95%CI were calculated accordingly. Statistical significance was defined as P < 0.05 (two-tailed).

The study protocol was approved by the Institutional Review Board of Southeast University Zhongda Hospital (Approval No. 2016ZDKYSB017), and the written informed consent was obtained.

RESULTS

Two hundred EPL cases aged 63.01 ± 6.08 years and 200 healthy controls aged 62.85 ± 6.03 years were enrolled. Among the pairs, 100 were males and 100 were females. Table 1 shows that 47 (23.5%) and 58 (29.0%) out of 200 cases and 200 controls, respectively, had H. pylori infection. Two independent samples t-test and conditional logistic regression analysis indicated that there were no statistically significant differences in age, BMI, education level, annual income per person, current drinking status, or H. pylori infection between the two groups after adjustment for covariates (P > 0.05). Compared with non-smokers, a smoking habit of more than 20 cigarettes a day was significantly associated with an elevated risk of EPL (P < 0.05).

Table 1 Characteristics and potential factors in cases with esophageal precancerous lesions and controls.

Category	Cases, n = 200	Controls, n = 200	Adjusted OR (95%CI)1	P value
Age (yr), mean ± SD	63.01 ± 6.08	62.85 ± 6.03		0.792a
BMI (kg/m2), mean ± SD	24.52 ± 3.33	24.36 ± 3.37		0.631a
Normal (18.5-23.9)	82 (41.0%)	84 (42.0%)	1.00 (reference)
Underweight (< 18.5)	4 (2.0%)	5 (2.5%)	0.61 (0.12-3.02)	0.545
Overweight (24.0-28.0)	84 (42.0%)	89 (44.5%)	0.95 (0.61-1.49)	0.836
Obese (> 28.0)	30 (15.0%)	22 (11.0%)	1.61 (0.83-3.12)	0.164
Education level
Illiterate	100 (50.0%)	96 (48.0%)	1.00 (reference)
Primary school education	74 (37.0%)	77 (38.5%)	0.79 (0.45-1.39)	0.413
Middle school education and higher	26 (13.0%)	27 (13.5%)	0.81 (0.36-1.79)	0.599
Annual income/person (RMB)
1-5000	53 (26.5%)	42 (21.0%)	1.00 (reference)
5001-10000	88 (44.0%)	86 (43.0%)	0.75 (0.45-1.25)	0.267
> 10000	59 (29.5%)	72 (36.0%)	0.67 (0.37-1.21)	0.183
Current smoking status (number of cigarettes/d)
Non-smoker	126 (63.0%)	134 (67.0%)	1.00 (reference)
1-10	20 (10.0%)	17 (8.5%)	1.39 (0.67-2.87)	0.381
11-20	38 (19.0%)	41 (20.5%)	1.10 (0.61-1.97)	0.755
> 20	16 (8.0%)	8 (4.0%)	3.11 (1.00-9.63)	0.049
Current drinking status (alcohol units consumed/d, 1 unit is 8 g or 10 mL of pure alcohol)
Non-drinker	147 (73.5%)	151 (75.5%)	1.00 (reference)
< 4	10 (5.0%)	10 (5.0%)	1.03 (0.41-2.59)	0.954
4-	26 (13.0%)	23 (11.5%)	1.02 (0.52-2.02)	0.946
8-	17 (8.5%)	16 (8.0%)	1.06 (0.47-2.42)	0.885
H. pylori infection
Negative	153 (76.5%)	142 (71.0%)	1.00 (reference)
Positive	47 (23.5%)	58 (29.0%)	0.75 (0.46-1.24)	0.265

^aP value of two independent samples t-test.

¹Conditional logistic regression model with adjustment for gender, age, BMI, education level, annual income, number of cigarettes per day, and alcohol units consumed per day, except the specific variable itself.

H. pylori: Helicobacter pylori; BMI: Body mass index; OR: Odds ratio.

Based on routine endoscopy examination, the study found that the number of cases whose EPL developed in upper, mid, and lower thoracic esophagus was 3, 130, and 67, respectively. Table 2 shows that the control group had the highest positive rate of H. pylori infection (29.0%), followed by EPL cases of upper and mid thoracic esophagus (24.8%) and EPL cases of lower thoracic esophagus (20.9%), but there was no statistically significant differences.

Table 2 Difference in Helicobacter pylori infection among controls and cases with esophageal precancerous lesions.

Group	H. pylori infection			P valuea
	Negative	Positive	Positive rate
Controls	142	58	29.0%
EPL cases (upper and mid thoracic esophagus)	100	33	24.8%	0.384
EPL cases (lower thoracic esophagus)	53	14	20.9%

^aP value of Fisher’s exact test.

H. pylori: Helicobacter pylori.

As shown in Table 3, when subjects were dichotomized according to gender, age, and the use of tobacco and alcohol, the protective effect of H. pylori infection against the risk of EPL was found in the group of drinkers after adjustment for covariates (OR = 0.32, 95%CI: 0.11-0.95). Supplementary Tables 1 and 2 shows that there may be a nonsignificant decreasing trend of H. pylori infection rate when alcohol consumption is increasing.

Table 3 Association between Helicobacter pylori infection and esophageal precancerous lesion risk dichotomized by gender, age, cigarette smoking, and alcohol drinking.

	Cases	Controls	Crude OR (95%CI)	P value	Adjusted OR (95%CI)1	P value
Male	n = 100	n = 100
H. pylori (-)	78 (78.0%)	72 (72.0%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	22 (22.0%)	28 (28.0%)	0.73 (0.38-2.38)	0.328	0.64 (0.32-1.27)	0.200
Female	n = 100	n = 100
H. pylori (-)	75 (75.0%)	70 (70.0%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	25 (25.0%)	30 (30.0%)	0.78 (0.42-1.45)	0.429	0.82 (0.42-1.58)	0.548
Age < 65 years	n = 107	n = 107
H. pylori (-)	76 (71.0%)	73 (68.2%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	31 (29.0%)	34 (31.8%)	0.88 (0.49-1.57)	0.656	0.89 (0.47-1.67)	0.708
Age ≥ 65 years	n = 93	n = 93
H. pylori (-)	77 (82.8%)	69 (74.2%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	16 (17.2%)	24 (25.8%)	0.60 (0.29-1.22)	0.156	0.59 (0.27-1.28)	0.183
Cigarette smoking (-)	n = 126	n = 134
H. pylori (-)	97 (77.0%)	93 (69.4%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	29 (23.0%)	41 (30.6%)	0.68 (0.39-1.18)	0.170	0.74 (0.42-1.32)	0.310
Cigarette smoking (+)	n = 74	n = 66
H. pylori (-)	56 (75.7%)	49 (74.2%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	18 (24.3%)	17 (25.8%)	0.93 (0.43-1.99)	0.845	0.80 (0.34-1.86)	0.601
Alcohol drinking (-)	n = 147	n = 151
H. pylori (-)	109 (74.1%)	109 (72.2%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	38 (25.9%)	42 (27.8%)	0.91 (0.54-1.51)	0.702	0.94 (0.55-1.61)	0.831
Alcohol drinking (+)	n = 53	n = 49
H. pylori (-)	44 (83.0%)	33 (67.3%)	1.00 (reference)	-	1.00 (reference)	-
H. pylori (+)	9 (17.0%)	16 (32.7%)	0.42 (0.17-1.07)	0.070	0.32 (0.11-0.95)	0.040

¹Adjustment for gender, age, BMI, education level, annual income, number of cigarettes per day, and alcohol units consumed per day.

H. pylori: Helicobacter pylori; 95%CI: 95% confidence interval; OR: Odds ratio.

Figure 2 illustrates the association between dietary factors and the risk of H. pylori infection after the adjustment for covariates via the unconditional logistic regression model. The result indicated that peanut intake was significantly associated with a reduced risk of H. pylori infection (OR = 0.39, 95%CI 0.20-0.74). Supplementary Table 3 shows that there may be a significant positive association between peanut consumption and alcohol drinking (P for trend < 0.05).

Figure 2 Association between dietary intake and the risk of Helicobacter pylori infection after adjustment for covariates via unconditional logistic regression model. BMI: Body mass index; OR: Odds ratio.

DISCUSSION

This study revealed that in drinkers, there was an association between H. pylori infection and a reduced risk of EPL, which is an identified early stage of esophageal carcinogenesis. However, the relationship between H. pylori infection and ESCC is still subject to much discussion. Some researchers believed that infection with H. pylori can increase the risk of ESCC by causing gastric atrophy that promotes excessive bacterial growth and causes endogenous nitrosamine production[24-26]. However, other studies which held that H. pylori infection probably plays a protective role in ESCC postulated that the protection is mediated via gastric atrophy, whereas the mechanism is related to a reduced load of esophageal acid[27,28]. Therefore, it is likely that ESCC might be affected in a double-edged manner by H. pylori infection, which is dependent on population and other possible external factors. For example, previous studies have indicated that acid regurgitation may be facilitated by the reduction in lower esophageal sphincter’s pressure and the retard of both esophageal motility and gastric emptying due to large consumption of alcoholic beverages[29-33]. Therefore, the current hypothesis is that H. pylori infection just alleviates esophageal reflux caused by alcohol to some extent, thus reducing the risk of esophageal carcinogenesis caused by acid reflux. Our results also reported that EPL cases of lower thoracic esophagus had the lowest positive rate of H. pylori infection, which may support the hypotheses to some extent, although the difference was not statistically significant. In addition, there is more data indicating the positive role of this bacterium for humans. For example, a recent review considered the data on H. pylori and suggested that H. pylori may be a latent or opportunistic pathogen rather than a true pathogen of some diseases, and is possibly part of the normal human microbiome as a commensal or even a symbiont organism[34]. However, it was reported that a regular but moderate alcohol intake could possibly facilitate elimination of H. pylori infection[35]. Supplementary material also shows a nonsignificant decreasing trend of H. pylori infection rate when alcohol consumption is increasing. This partly supports the hypothesis that there is a possibility that drinkers without H. pylori infection could have more alcohol consumption. In other words, the decreased EPL risk in drinkers with H. pylori infection is possibly related to a lower alcohol consumption. However, because the result was not statistically significant, and there was no significant association between alcohol consumption and EPL risk in Huai’an in both this study and the previous epidemiological investigation[21], it is hard to address whether the reduced risk of EPL in drinkers with H. pylori infection was related to a reduced alcohol intake.

Additionally, the present study reported that the consumption of peanuts may provide protection from H. pylori infection. Since peanuts are high in fat, the duodenal mucosa secretes the hormone enterogastrone when fatty food is present in the stomach or small intestine[36]. Enterogastrone inhibits gastric movements and secretion of gastric acid, possibly by blocking the production or activity of gastrin, the hormone that initially leads to these functions[37]. Therefore, the reduced amount of acid produced may influence the growth of H. pylori, as H. pylori is dependent on acidity to survive for a long time[38]. In addition, in China, people are likely to drink and eat peanuts at the same time, and Supplementary material shows that there was a positive association between peanut consumption and alcohol drinking. Therefore, the inverse association between the consumption of peanut and the risk of H. pylori infection may be mediated by alcohol drinking. However, there is still a lack of evidence to verify the above hypotheses, thus further researches are required to evaluate the relationship between peanut consumption and H. pylori infection.

At present, about 50% of the global population and more than 70% of the population in some developing countries are infected by H. pylori[39]. However, this study reported that the positive rates of H. pylori infection were only 23.5% and 29.0% in EPL cases and healthy controls. In an early study conducted by Gao et al[40], Huai’an, Jiangsu Province was selected as a high incidence area of upper digestive tract cancers, and Pizhou, Jiangsu Province was selected as a low incidence area. They used ELISA and latex agglutinate test for the detection of H. pylori infection, and found that the prevalence of H. pylori infection among the gastric cancer group/upper digestive tract cancer group in the low incidence area of Pizhou (66.67%/63.46%) was significantly higher than that in the high incidence area of Huai’an (38.64%/39.33%). However, in the high incidence area of Huai’an, the prevalence of H. pylori infection in non-cancer controls and the healthy family members of the cancer cases was higher than that of cases. Therefore, the previous study and our study found that the prevalence of H. pylori infection in Huai’an may be much lower than that in other areas, and the prevalence in upper digestive tract cancers or EPL cases can be lower than that in non-cancer population in this region.

CONCLUSION

In summary, our study suggested that H. pylori infection is likely to decrease EPL risk in drinkers for a rural adult Chinese population, and the consumption of peanuts may be related to a reduced risk of H. pylori infection. However, the sample size used is a limitation of the study, which may bring difficulties to evaluate statistical significance in some statistical analyses, thus the findings should be framed as preliminary evidence. A case-control study might be difficult to determine causality, so the statement of “protective role” might be overestimated. Hence, it is necessary to design a large-scale prospective cohort study to address the impact of H. pylori infection on ESCC, the localization of lesions, and the association with dietary intake and the use of alcohol in the future. Additionally, the low prevalence of H. pylori infection in Huai’an is a peculiar finding, which implies that further investigations are recommended.

ARTICLE HIGHLIGHTS

Research background

The role of Helicobacter pylori (H. pylori) infection in esophageal squamous cell carcinoma (ESCC) remains a topic of much debate.

Research motivation

To assess the relationship between H. pylori infection and the risk of precancerous lesions of ESCC, which is an identified early stage of carcinogenesis.

Research objectives

This study aimed to evaluate the association between H. pylori infection and the risk of esophageal precancerous lesions (EPL) in a high-incidence area in Huai’an, and further explore the association between dietary factors and the risk of H. pylori infection.

Research methods

The study was based on a case-control design. Epidemiological data were collected and H. pylori seropositivity was tested. An unconditional logistic regression model was used to analyze the association between H. pylori infection and EPL risk with adjustment for confounders, as well as the association between dietary factors and risk of H. pylori infection.

Research results

The control group had the highest positive rate of H. pylori infection (29.0%), followed by EPL cases of upper and mid thoracic esophagus (24.8%) and EPL cases of lower thoracic esophagus (20.9%). The protective effect of H. pylori infection against the risk of EPL was observed in the group of drinkers after adjustment for covariates [odds ratio (OR) = 0.32, 95% confidence interval (95%CI): 0.11-0.95]. Peanut intake was significantly associated with a reduced risk of H. pylori infection (OR = 0.39, 95%CI: 0.20-0.74).

Research conclusions

H. pylori infection may decrease the risk of EPL in drinkers for a rural adult Chinese population, and the consumption of peanuts may be related to a reduced risk of H. pylori infection.

Research perspectives

A well-designed prospective cohort study is required to address the impact of H. pylori infection on ESCC, the localization of lesions, and the association with dietary intake and alcohol drinking. Additionally, the low prevalence of H. pylori infection in Huai’an is a peculiar finding, which implies that further investigations are recommended.

ACKNOWLEDGEMENTS

Thanks to all the subjects and researchers for their contributions and hard work. We really appreciate Mr Rob Unwin’s assistance in polishing the manuscript.