Epidemiology of Helicobacter pylori Infection in Tay Children in Vietnam

Background: Differences in prevalence of H. pylori among racial and ethnic groups have been described worldwide. The aim of the present study was to evaluate the seroprevalence of H. pylori infection and factors associated with the H. pylori infection among Tay children and adults living in Vietnam. Materials and methods: In a cross-sectional study 1,094 healthy individuals of all generations living together in the same home were selected from 278 households, based on cluster sampling of residential location between September and December 2013 in Langson province. H. pylori infection status was determined by serology test on samples obtained at each visit. A questionnaire was filled out at the start of the study. H. pylori serology data were analyzed using χ2 test and logistic regression models. Results: An overall H. pylori seroprevalence was 46.8%. H. pylori seroprevalence was 51.4% in adults versus 41.4% in children ≤18 years old (p<0.05). Regular hand washing after defecation and breastfeeding over 12 months were protective factors for H. pylori infection [OR (95%CI): 0.73, (0.39-0.92); 0.71 (0.35-0.94); respectively]. H. pylori infected mothers, first siblings and grandparents were found as risk factors for H. pylori infection in children [OR (95%CI): 2.98 (1.13-3.14), 1.4 (1.0-3.12), 1.41 (1.11-4.64); respectively]. No other factors such as size of household or sibling, infected fathers, regular sharing bed, collective life initiation and antibiotic use were found to be significant risk factors for infection Conclusion: The first community-based study in Tay population showed moderate and indifferent rates of H. pylori infection. Data from the present study are consistent with intra-familial H. pylori transmission and suggest that improvement of living conditions should be protective against infection.


Introduction
Helicobacter pylori (H. pylori) is one of the most common infections in humans, with an estimated 50% of the world's population being infected [1]. This organism has been implicated in the pathogenesis of active and chronic gastritis, peptic ulcer and gastric cancer [1,2]. Childhood has been identified as the critical time for acquisition of H. pylori infection [3][4][5][6]. The exact route of transmission remains unclear but the infection clusters in families and familial spread is thought to be the major mode of transmission, both in developed and developing countries [7]. The prevalence of H. pylori infection in children all over the world is diverse and dependent on many factors. The highest rates H. pylori infection are reported in developing countries with lower socioeconomic and hygiene conditions as well as a high density of people in the household, while the lower percentage of infected children is observed in high income countries. Differences in prevalence among racial and ethnic groups have been described worldwide, but it is unclear to what extent such differences can be ascribed to socioeconomic factors and other possible risk factors [4,8]. Vietnam is the easternmost country in Southeast Asia with an estimated 90.5 million inhabitants. Vietnam is also home to 54 ethnic groups with different cultures, of which 75% to 80% living in rural or remote areas [9,10]. High rates of H. pylori infection was reported in both hospital-based and populationbased studies. Prevalence of infection with H. pylori has been reported between 50% and 80% in several studies conducted in adults and around 26.0% to 71.4% in children [11][12][13]. Different risk factors for has been identified H. pylori infection in Kinh ethnic majority [11][12][13] and in some among other ethnic minorities [14,15]. As the socioeconomic level and lifestyle vary considerably among ethnic groups of the country, the prevalence of H. pylori infection and particularly risk factors for this infection were different in previous studies [11][12][13][14][15]. According to the general statistics office of the government of Vietnam 2009 [10], the Tay is the second largest ethnic group in Vietnam after the majority Kinh ethnic group with an estimated 1.7 million people. Tay people have their own habits, customs, cultural practices and economic condition that differ from those of Kinh people [10]. No data is available about prevalence of H. pylori infection and risk factors for this infection in Tay people to develop appropriate preventative measures for managing the H. pylori infection in this subpopulation in the future. The aim of the present study was to evaluate the seroprevalence of H. pylori infection and factors associated with the H. pylori infection among Tay children and adults living in the Langson province.

Study population
Langson is a province in far northern Vietnam, covers an area of 8327.6 square kilometers and a population of 759,000 people. Tay people comprised 35.92% of the population [9,10]. Bacson is a district of Langson with population of 65,073 people over a total area of 698 square kilometers where we implemented the study, enrolling 1,094 members of all generations living together in the same home from 278 households.
To avoid a selection bias, children aged less than 6 months were excluded due to the likelihood of residual maternal H. pylori antibody, as well as those with severe diseases or immuno-compromised status due to the possibility of an altered immune response.
Data were collected after obtaining written consents from local administration and health authorities. Informed consent was obtained from each household member. The study was granted ethical clearance in Vietnam by the ethics committee of the Hanoi Medical University.

Data collection
In this cross-sectional study, door-to-door sampling method was adopted. We used a structured questionnaire, as in previous studies [12][13][14][15] for data collection on sociodemographic, health status and potential exposure. The questionnaires were completed by the study investigators and information was collected from the head of the households. Blood samples (5 mL) were obtained from each subject by peripheral venipuncture under aseptic conditions and were immediately centrifuged. Sera were separated and preserved in vaccine thermos, then sent to the reference laboratory (Microbiology Division of Digestive Diseases, National Institute of Epidemiology and Hygiene) on the same day where sera were stored at -20°C and processed as previously described [16].

Variable definitions
H. pylori infection was determined by in-house ELISA for dosage of H. pylori IgG an body against specific H. pylori antigen. The ELISA was carried out in the reference laboratory (Microbiology Division of Digestive Diseases, National Institute of Epidemiology and Hygiene) using sonicated Swedish and Vietnamese H. pylori strains as antigens prepared and validated in the microbiology department, Karolinska Institute, Stockholm for use in Vietnamese adults and children (sensitivity of 99.6%, and specificity of 97.8%) [16]. Children's sera with a value equal to or over 0.18 optical density (OD) unit were classified positive.
Socio-demographic variables consisted of age and sex of all children and adults in every household, monthly family income classified into three categories (i.e. <500.10 3 VND, actual national minimal monthly income, 500 to 1,000 or >1,000.10 3 VND), household space (in 3 categories, i.e. up to 10 m 2 /capita according to national mean standard, 10 to 20 or >20 m 2 / capita), parents' occupation and parents' education level [10].
Potential exposure variables were divided into three major subgroups Environmental, individual hygiene and life-style including water sources classified in 2 main sources (family well and others such as streams, raining or collective wellness), latrine in 2 main types (existence or not), pet (dog or cat) in the house in 2 categories (presence or not), behavior of children on hand washing before meal and after toilet in 2 categories (not regular for those who practiced less frequently than one per twice, and regular for those who practiced every time or every of other time), mouth-to-mouth feeding from mothers or caregivers to child in 2 categories (regular or not), breast feeding duration in 2 categories (shorter or longer than 12 months of age).
Health status variables consisted of H. pylori infection status in parents and sibling (H. pylori seropositivity or seronegativity), gastro-duodenal history of child and parents (defined as having had a gastro-duodenal disease if this disease was diagnosed and treated by health-givers from district health center or higher levels) in 2 categories (presence or absence), and history of child's antibiotic use during the previous 6, 12 months determined from individual health booklets or in-depth interviews of the main caregivers, in 2 categories (absence or presence).

Statistical analysis
First, study population characteristics were compared according to their H. pylori infection status using the Chi square (χ 2 ) test. As one of the major aims of our study was to detect the factors potentially influencing H. pylori infection in this subpopulation, the appropriate strategy of analysis must be able to control for mutual confounding, point out the risk factors and avoid overlooking important associated variables. We then analyzed separately the associations between H. pylori infection status with the demographic and socio-economic variables and with variables related to potential exposure. Analysis was performed firstly by univariate technique, by adjusting in each group on every variable, and finally by using backward stepwise conditional logistic regression to select variables importantly associated with H. pylori infection within each group, including all variables significantly associated with H. pylori seropositivity after adjustment and those with p values less than 0.20 by Chi square test. Associations were expressed as odds ratio (OR) and their confidence intervals (95% CI). Finally, backward stepwise conditional procedures were used again to include in the final model not only variables independently associated with H. pylori serological status in each group, but also those known to be important for transmission pathways. Statistical significance was set up at the 0.05 level. All p values were 2tailed. Data were analyzed using SPSS software (SPSS® for Windows TM version 16.0 Copyright SPSS Inc.).

Results
Among 1,094 individuals enrolled in the study, 512 (46.8%) were seropositive for H. pylori infection. H. pylori seroprevalence was 51.4% (304/591) in adults and 41.4% (197/476) in children ≤18 years old (p<0.05). Rates of H. pylori infection is associated with socio-demographic variables in Tay children aged 6 months to 18 years is presented in Table 1. There were no significant differences in H. pylori seropositivity based on sex (p>0.05). The prevalence of H. pylori infection was 30.9% for children under 3 years of age, rising to 53.1% for those older than 15 years. When analyzed by age strata, the odds of infection were higher for children aged over 15 years compared with those aged less than 3 years (odds ratio [OR], 1.59; 95% CI, 1.09-2.32). No significant relationship was found between H. pylori seropositivity and variables related to socioeconomic status of their household. H. pylori seropositive in Tay children in relation to variables related to promiscuity and health status of study population as well as antibiotic use in children are presented in Table 3. The prevalence of H. pylori was higher among children with a family member infected with the bacterium than among children without any infected family member. The differences were significant for mothers, siblings, and also grandparents (p<0.001). We did not find any relationship between H. pylori seropositivity in children and variables related to promiscuity (i.e. size of household or sibling, H. pylori infection in fathers, regular sharing bed, collective life initiation and history antibiotic use within 6 or 12 months in children). In the final model of multivariate analysis (data not shown), low risk for H. pylori infection in children was independently associated with regular washing hand after defecation (OR 0.73 (0.39-0.92). In contrast, H. pylori infection in mothers, first sibling and grandparents were found associated with higher risk factors for H. pylori infection in children (OR (95%CI): 2.98 (1.13-3.14); 1.4 (1.0-3.12); 1.41 (1.11-4.64); respectively).

Discussion
Our study showed an overall prevalence of H. pylori infection in Tay people was 46.8%. H. pylori seroprevalence was 51.4% in adults and 41.4% in children ≤18 years old. The seroprevalence of H. pylori infection of the present study was comparable to those reported in other community-based studies in Vietnam [11][12][13][14][15]. Hoang et al. conducted a community-based study in Kinh majority ethnic in Hanoi showed an overall seroprevalence of H. pylori of 58.2% with 47.3% in children [11]. In another community-based study in Kinh of a rural area from Nghean province, Nguyen et al.
reported an overall rate of H. pylori infection of 61.5% and 55.5% prevalence in children [17]. Recently, Le et al. reported an overall rate of 45.2% with 40.0% in children under 15 in their community-based study carried out in a population of Kinh and 4 other minority ethnics in central highland [15]. However, the overall H. pylori infection rate and rate of infection in children in the present study was comparable to the higher than those of the previous study carried out in a population of Kinh and 3 other minority ethnics in a mountainous area conducted by Trinh et al. [14].
It is believed that H. pylori is mainly acquired during childhood and little is known about its age of onset, rate, or mode of colonization [3][4][5]. Okuda et al. followed-up 108 out of 237 Japanese infants until 24 months by a fecal antigen test found that 16 infants turned to be HpSA positive within 12 months. Among them four infants remained positive by the consecutive tests demonstrated infants seemed to acquire H. pylori infection in the first year of life [18] [12]. Our data is thus in line with the finding of others studies presently.
A significant association of H. pylori with male was showed in some previous studies [21]. The seroprevalence of H. pylori was significantly higher in boys (p<0.01) in two studies conducted in Spain and Japan. In contrast, the rate of H. pylori infection was not found to be related to gender in Taiwan, Korea, and Mexico [22][23][24]. There was no significant difference between the prevalence of H. pylori infection and gender in our study.
Low socioeconomic status and poor sanitary standards have been described as risk factors for the acquisition and transmission of H. pylori [25]. No significant relationship was found between H. pylori seropositivity and variables related to socio-economic status of their household in our study. Data from some previous studies have revealed that socioeconomic level is a risk factor for H. pylori infection. The prevalence by age decreasing along with socioeconomic development was reported in a multi-center study from Czech Republic [26]. The rate of H. pylori infection was significant decreased from 41.7% in 2001 to 23.5% in 2011. Decrease in the rate of H. pylori infection in children was also reported in Estonia [27] and Russia [28]. It has been explained by improving socioeconomic conditions and standards of living together. Mother's educational levels rather than the fathers were associated with the rate of H. pylori infection in children has been shown in Argentina study [29]. One Korean study showed that the prevalence of H. pylori infection was lower in children from families of higher socioeconomic status [23]. The poor socioeconomic status of people in southwest Tehran is correlated with a higher incidence of infection (p<0.05) [30].
Potential exposure variables in relation to individual hygiene as well as exclusive habits and life-style have been always sought by investigators in epidemiological studies on H. pylori infection [22,23]. The present study showed no significant differences in rates of H. pylori infection based on water source, latrine presence, regular hand washing before meal and having pets. Children doing hand washing after defecation was less likely to H. pylori seropositivity (OR: 0.63, 95%CI: 0.49-0.96, p=0.023). In the present study, Nguyen et al. has shown that owing latrine in the family and regular washing hand before meal is significantly associated to lower H. pylori seroposivity [13,31]. In contrast, regular washing hand after defecation and no taking food by hand appeared as protective factors for H. pylori infection in ethnic people in Vietnam. Abebaw et al. also reported higher seroprevalence of H. pylori in people who used unprotected surface water and irregular washing hand before meal [32].
Breastfeeding more than 12 months was negatively and independently associated with H. pylori seropositivity in the present study (adjusted OR 0.71; 95% CI: 0.35-0.94, p=0.036). Some studies have shown breastfeeding's protective effect against the acquisition of H. pylori infection [33,34], while other studies have reported that breastfeeding does not have a protective effect against the acquisition of H. pylori infection [35,36]. In a case-control study parents of children with and without H. pylori infection who had undergone endoscopic survey and gastric biopsy in the Tehran, Monajemzadeh et al. found that breastfeeding in the first 6 months after birth may decrease the degree of H. pylori colonization, postpone infection until older age, shorten the duration of symptoms, and be concomitant with milder gastritis [37]. Results from previous study showed that children breastfed longer than 12 months were likely to be more infected by H. pylori than those breastfed for a shorter duration (OR: 1.64; 95% CI: 1.09-2.71) [30]. A systematic review conducted by Chak showed that breast-feeding is protective against H. pylori infection (OR 0.78; 95% CI: 0.61-0.99; p=0.02) [34]. OR was 0.55 (95% CI: 0.33-0.93; p=0.01) in studies in which the subjects resided in developing countries compared with 0.93 (95% CI: 0.73-1.19; p=0.28) in those of developed nations. Thomas et al. found a positive relation between the age at acquisition of the infection and amount of anti-H. pylori IgA in human breast milk in Gambian children [38]. They concluded that specific human milk IgA may have a crucial role in delaying the onset of H. pylori infection. Gold et al. showed an association between breastfeeding and an increased risk for H. pylori IgGseroconversion during 14 months of follow-up. The possible mechanisms enrolled in this protection may be the lactoferrin in human milk which binds to H. pylori liposaccharide inactivating the organism [39] and breast milk inhibited H. pylori adherence to gastric adenocarcinoma cell line in vitro [40].
The prevalence of H. pylori was higher among children with a family member infected with the bacterium than among children without any infected family member. Our study showed the presence of infected siblings, mother and grandparents as a risk factor for the infection in children. We did not find any relationship between H. pylori seropositivity in children and variables related to promiscuity such as size of household or sibling, H. pylori infection in fathers, regular bed sharing, collective life initiation and history of antibiotic use within 6 or 12 months in children. Some authors have demonstrated the mode of H. pylori transmission was infection clusters in families and familial spread [35,41]. Data from those studies provided an evidence that having infected family members is highly associated with the infection in children. Drumm et al. reported seroprevalence of 80% in children whose has infected siblings compared to 13% of age-matched controls [41]. Malaty et al. also provided data demonstrating an increased prevalence of colonized children of infected parents [42,43]. The higher prevalence of infection due to H. pylori in parents of infected children suggests person-toperson transmission within the family [22,35]. The H. pylori status of the mother was found to be a strong determinant for childhood infection and more predictive than the status of the father. Kivi et al. showed that having an infected mother (OR: 11.6; 95% CI: 2.0-67.9) or at least one infected sibling (OR: 8.1, 95% CI: 1.8-37.3) was a major risk factor for H. pylori infection in Swedish children [41]. The mother is likely to have introduced the infection to her offspring, because daycare attendance and H. pylori prevalence amongst classmates have been refuted as risk factors for the infection in this child population, thus rendering child-to-child transmission outside the family improbable in Sweden [44]. Goodman et al. showed that having infected siblings was a predictor of the infection in children, although that study did not control for parental infection status [45]. The importance of both infected siblings and mothers was recently corroborated in a Brazilian highprevalence community [46]. The findings are substantiated by a previous report of H. pylori strain concordance between mothers and offspring and amongst siblings, demonstrated by using bacterial molecular typing in a subset of the currently studied families [47]. Aguemon et al. also reported the higher infection rate in children whose parents were both infected or only infected mother. By using logistic regression analysis, sharing bed (OR: 3.85; 95% CI: 1.53-9.67, p<0.003) and infected mothers (OR: 9.82, 95% CI: 4.13-23.31, p<0.001) were independent predictors for H. pylori infection. Family contact with infected persons and crowded living conditions were associated with increased risk of infection [4,22,43]. In the present study, we observed rates of higher positivity among children of H. pylori positive mothers or grandparents compared to those whose family members were negative. Result from our study is as same as those from another study conducted in Mekong delta by Nguyen 1-3.3), p=0.001) were identified as risk factors for infection in H. pylori children [30]. The fact that H. pylori infection in siblings influenced more on the infection of other children than the infection in the mother might be explained by longer and closer contact between siblings than with mother, because in Vietnam, like in most developing countries, after the first year of age, children spent more time to play and to sleep together with siblings than with mother or father. Mouth secretions of mother could be contaminated with H. pylori may be transmitted to the infant and child. Transmission may occur also for sharing common cups, spoons, chopsticks, teats of feeding bottles, or for chewing or tasting children's food.
Strength of our study lies in this cross-sectional community based-study. We recruited a large population who were all members of all generations living together in the same household. This method of sampling facilitated the data analysis and interpretation, provided a more comprehensive understanding about the interaction or interrelation between studied variables and H. pylori infection. Therefore, it rendered more feasible and more reliable for risk factors to be identified among studied socio-demographic and potential exposure variables. Because of the possible limitations of our study in our study are issued from the limitation of structured questionnaire reluctantly adopted according to human and financial conditions of the study. The first limitation resides in recall biases inevitably committed by respondents during interview. One more limitation is ambiguity and difficulty to calculate exact income per capita per month given diversity in homemade and self-serving products, very popular in this population living in remote area of mountain province. Another point of limitation issued from limited knowledge of participants in recognizing history of gastro duodenal disorders and loss of health booklet or medical prescription in many cases, sources of inaccuracy in deciding whether someone had suffered from digestive disease in relation to H. pylori infection.

Conclusion
This first community-based study of epidemiology of H. pylori infection in the Tay population settled in mountain province showed moderate and indifferent rates of seropositivity of H. pylori infection. Data from the present study are consistent with intrafamilial H. pylori transmission and suggest that improvement of living conditions should be protective against infection.