Helicobacter pylori infection was examined among 356 asymptomatic white Hispanic and black children aged 2–16 years attending 13 licensed day care centers in Houston. Demographic information and socioeconomic factors were evaluated. H. pylori status was determined by 13C-urea breath testing. The prevalence of active H. pylori infection was 24% and increased with age. Prevalence was almost identical among white Hispanic and black children. Children living in the most crowded conditions were at the greatest risk for H. pylori acquisition, and an inverse correlation was seen between the mother's education and H. pylori positivity in children. Breast-feeding played a protective role against the acquisition of H. pylori infection. Understanding the epidemiology of H. pylori infection in childhood requires better understanding of the interactions between environment, ethnic group, and socioeconomic conditions.
It is now recognized that gastritis, gastric ulcers, duodenal ulcers, gastric carcinoma, and primary gastric B cell lymphoma are all associated with Helicobacter pylori infection [1–4]. H. pylori infection is known to cluster in families [5–8]; however, whether transmission occurs through a common exposure source or through contact between individuals remains unknown. The fact that the spouse of an index patient tends to have the same H. pylori status as the index patient suggests that environmental factors are at least as important as genetics in the transmission of H. pylori infection .
Young children have many behavioral characteristics that increase the risk for transmission of infectious diseases [9, 10], and childhood is known to be a time of high risk for H. pylori acquisition [11–14]. Nevertheless, the natural history of H. pylori infection remains poorly understood. A better understanding of the epidemiology of H. pylori infection in pediatric patients is required to understand the natural history of H. pylori infection and to identify the most common mode(s) of transmission, as well as how, when, and where to break the chain of transmission.
The high prevalence of H. pylori infection among white Hispanic and black adults has been shown to be inversely correlated with childhood socioeconomic status, suggesting that environmental factors are more important than genetic factors in determining the pattern of transmission . Because those results were derived from a distant birth cohort, we tested directly the prevalence of active H. pylori infection among minority children by means of the sensitive and specific 13C-urea breath test [15, 16]. The 13C-urea breath test is advantageous for use with children because it is non-invasive, valid, safe, and easy to perform, and, unlike serology, it yields positive results for recently acquired infection.
Enrollment. The study involved 13 licensed day care centers from different locations in Houston. Centers were chosen that primarily enrolled minority children from the low and lower-middle socioeconomic classes. Two of the studied centers were attached to family shelters, and the children's enrollment at the day care center was based on economic hardship in the family. These 2 centers were open 24 h daily to offer services to the family. In the other 11 centers that were enrolled in the study, the parents paid income-based fees for day care services. The target population was children between the ages of 2 and 16 years. Children >5 years old attended day care centers in the late afternoon, after regular school hours, and in the morning before school. Entry criteria included no acute or chronic stomach problems or gastric surgery and no current (within 1 month) use of antibiotics. The sampling of the study was not random but depended on invitation and eligibility as determined by the entry criteria.
A detailed verbal and written explanation of the study was given to each center director at the time of the site visit. General administrative information was obtained, and a letter explaining the study's objective, a consent form, and an instrument for gathering demographic and health information about the participating children was provided. After the center director's consent was obtained, the parents of children enrolled in day care were invited to attend presentations that introduced and explained the aims of the project. Presentations were given during an evening open house, to enable parents to question the investigators about the project. The study started in July 1996, and the phase reported in this article ended in October 1998.
Questionnaire. Mothers answered a detailed demographic questionnaire that included the child's age, sex, ethnic group, and place of birth and the occupation, income, and education level of the parent(s). Environmental information was collected regarding type of home, number of rooms and number of bedrooms in the home, number of children and adults living in the home, and breast-feeding practices. The questionnaire also included data regarding enrollment of the child in day care and the person financially responsible for payment to the center. In addition, information on whether the child had any kind of continuing stomach problem and whether the child had received antibiotics during the previous month was collected.
13C-Urea breath test. H. pylori status was assessed by the 13C-urea breath test, which indicates active infection [15, 16]. The 13C-urea breath test detects the presence of H. pylori from an enrichment of breath 13CO2, which is critically dependent on the amount of dilution by endogenous CO2 production. Typically the 13C-urea breath test has been scored on the basis of change in isotopic abundance from the pretest baseline value to the postdose value. This has sufficed for diagnostic purposes in adults. Because the labeled CO2 produced in the stomach by the organism combines with respiratory CO2, the degree to which the 13CO2 is diluted by host CO2 determines the degree of enrichment present in the expired breath. CO2 production differs in relation to age (greater in adults than in children), sex (greater in males than in females), weight, and height. For adults, the cutoff value of 2.4‰ equals 10 µg of urea hydrolyzed/min . It is possible to correct for size, age, and sex by use of a formula that incorporates the age, sex, height, and weight of each subject independently to determine the basal metabolic rate by means of the Schofield equations .
The 13C-urea breath test was administered by use of 120 mg of 13C-urea dissolved in 200 mL of white grape juice. Breath samples were collected before and 30 min after drinking the grape juice. A positive result was defined as an increase in the 13C-urea abundance of 10 µg of urea hydrolyzed/min .
Serological methods. Forty-five parents of the participating children gave their consent to have blood drawn from their children. Serum was separated and frozen until assay. IgG antibody to the high-molecular-weight cell-associated proteins of H. pylori was measured by an ELISA (HM-CAP; Enteric Products) that was validated for both adults and children [18, 19]. According to the manufacturer's instructions, a value >2.2 is considered positive for IgG antibodies to H. pylori.
Fifteen children tested positive for H. pylori by 13C-urea breath testing, and 30 children tested negative. ELISA testing was done blindly, without knowledge of the 13C-urea breath test results. Thirteen of the 15 children with positive results on 13C-urea breath testing had an ELISA value >3.0 for IgG antibodies. None of the children with negative 13C-urea breath test results had IgG antibodies, and there were no borderline results for the children with positive results of 13C-urea breath testing or ELISA.
Statistical analyses. Mantel-Haenszel χ2 analyses were done to measure the differences in the overall prevalence of the outcomes. Risk factors for H. pylori were analyzed by age-adjusted and multiple logistic regression, and weighted ORs and 95% CIs were calculated for all study variables. Sex and ethnic group were not included in the final model because of their negligible impact on the risk for acquiring H. pylori infection. Crowding status of the day care centers was categorized as high or low. High was defined as placement of all children, regardless of age, into 1 large undivided room (as was the case in 2 particular centers); low was defined as separation of children by age into classes or different areas where daily activities take place. All P values were 2-tailed; the α level of significance was set at .05. Data were analyzed by use of the SAS program (SAS Institute).
Prevalence of H. pylori infection in relation to age, ethnic group, and sex. The cohort consisted of 356 children, of whom 83 (23%) were white Hispanic and 273 (77%) black, and 162 (45.5%) were boys and 194 (54.5%) girls. The prevalence of active H. pylori infection increased from 13.5% for those 2–5 years old to 48.5% for those >10 years old (P = .002), with no significant difference between white Hispanic and black children (26.5% vs. 23%; P = .52) (figure 1). No significant difference was observed in H. pylori prevalence between boys (26%) and girls (22%) (P = .4).
Association between H. pylori infection and demographic, socioeconomic, and environmental indicators. Table 1 shows the crude and age-adjusted ORs for positive results of H. pylori13C-urea breath testing in relation to age, ethnic group, sex, socioeconomic indicators, type of day care center, and breast-feeding practices. Children attending the 2 most crowded day care centers were at greater risk for H. pylori acquisition (age-adjusted OR, 3.0; 95% CI, 2.2–6.1; P < .01) than were children attending the less crowded centers. The crude OR was 1.6 (95% CI, 0.9–3.5) for children who lived in 1-bedroom homes, compared with children whose homes had ⩾3 bedrooms; however, the magnitude of that OR was not significant, and it was further reduced when adjusted for age and other variables. Although H. pylori prevalence was higher for children in households with ⩾4 children than for those in households with ⩽3 children (age-adjusted OR, 2.4; 95% CI, 1.1–5.1), children living with ⩾3 adults had a lower prevalence (11%) than did those living with 1 or 2 adults (30%; P < .05).
There was a statistically significant inverse trend for the association between H. pylori acquisition and the mother's education level. Children whose mothers did not complete high school had a higher rate of H. pylori infection than did children whose mothers had a partial or complete college education (43% versus 8%, respectively; P < .001). This difference remained statistically significant after adjusting for the child's age (age-adjusted OR, 8.5; 95% CI, 2.5–14.2). Breast-feeding practices had a significant protective effect against acquisition of H. pylori infection, and the effect remained before and after adjusting for age (age-adjusted OR, 3.0; 95% CI, 1.5–5.9).
In a multivariate model, we included age, number of children and adults in the household, number of bedrooms in the home, mother's education level, type of day care center attended, and breast-feeding practices. After multivariate adjustment, patterns for age, mother's education level, type of day care center attended, and breast-feeding practices were statistically significant at P < .05. However, the findings for household structure, including number of adults and children in the household and number of bedrooms in the home, were not statistically significant, at P < .05, when associated with H. pylori acquisition.
Cross-sectional studies have consistently shown a gradual increase in H. pylori seroprevalence with age [20–23], which has been interpreted as a birth cohort effect, reflecting a fall in the rate of H. pylori acquisition in successive generations of children as sanitation improved and standards of living increased . Although it is commonly believed that the prevalence of H. pylori infection among children in developed countries is low, the current study found that 24% of nonaffluent white Hispanic and black children who were born in the United States had active H. pylori infection. Of particular interest was the finding that, by the age of 10 years, 20% of the children were infected with H. pylori. Whereas the findings of the current study are consistent with our previous study, in which we reported a high prevalence of H. pylori infection among white Hispanic and black adults , they contrast with the low prevalences reported among children from European countries, which range from 5% to 10% [26–28]. These results show that, even in developed countries, there are large populations of native-born children among whom H. pylori infection is common. Studies from the United States have reported seroprevalence rates of H. pylori infection in the pediatric population of 12%–44% [29–31]. This wide variation in the prevalence of H. pylori infection is likely due to the wide diversity of socioeconomic status, ethnic groups, and social and cultural backgrounds.
The majority of the children <6 years old who were studied spent at least 8 h daily, 5 days a week, at the day care center, whereas older children divided their day, attending the day care center before or after school. The finding of a significant relationship between the crowded conditions in the day care center and H. pylori acquisition among the children studied highlights the importance of factors outside the home for transmission of H. pylori infection. These findings are consistent with a study from Edinburgh that reported that attending school in a socially deprived area was an independent risk factor for children to acquire H. pylori infection .
Although the literature on H. pylori and breast-feeding practices is limited, a few studies have hinted at the protective effect of breast milk against acquisition of H. pylori infection [33, 34]. A prospective study of infants from a Gambian village found that a high level of IgA in the mother's breast milk might protect against infection during the first year of a child's life . The mechanism by which human milk might reduce the rate of acquisition of H. pylori infection was examined in an in vitro study that evaluated the effect of human milk on the adherence of H. pylori to gastric cells . That study reported that human milk inhibits adherence of H. pylori to KATO III cells, independent of whether the donor was already infected with H. pylori. The results of our study provide additional evidence in support of an association between breast-feeding and prevention of H. pylori infection. Although breast-feeding was significantly associated with the mother's education level in this study, children who were breast-fed had a lower rate of H. pylori infection than did children who were not breast-fed, regardless of the mother's education (figure 2).
In conclusion, our study showed that the epidemiology of H. pylori infection in childhood might involve environments both inside and outside the home. The high prevalence of H. pylori infection among minority children in the United States raises several questions regarding treatment and management of the infection. When, where, and why a child becomes most vulnerable to the infection is still unknown. A subsequent, detailed study will be required to identify these critical periods and to develop and test strategies to prevent transmission of the infection.