Abstract
Objective
Helicobacter pylori (H. pylori) colonization rates in childhood have declined in Western populations, but it is unknown whether this trend is similar in children of non-Western ethnic backgrounds, born in a Western country. We aimed to identify H. pylori status in children, and determine both mother-to-child transmission and risk factors for colonization.
Design
Antibodies against H. pylori and cytotoxin-associated gene A (CagA) were measured in children participating in a population-based prospective cohort study in Rotterdam, the Netherlands. Information on demographics and characteristics was collected using questionnaires.
Results
We analysed the serum of 4,467 children (mean age 6.2 years ±0.5 SD) and compared the results with the H. pylori status of their mothers (available for 3,185 children). Overall, 438 (10%) children were H. pylori-positive, of whom 142 (32%) were CagA-positive. Independent risk factors for colonization were: maternal H. pylori positivity (OR 2.12; 95%CI 1.62–2.77), non-Dutch ethnicity (OR 2.05; 95%CI 1.54–2.73), female gender (OR 1.47; 95%CI 1.20–1.80), and lower maternal education level (OR 1.38; 95%CI 1.06–1.79). Comparing mothers and children, we found an intergenerational decrease of 76% and 77% for Hp+CagA− and Hp+CagA+-strains, respectively, consistent across all nine ethnic groups studied. Male gender, higher maternal educational level, and no older siblings, were independently associated with loss of H. pylori.
Conclusions
Although the highest H. pylori and CagA prevalence was found in children of non-Dutch ethnicities, the decreased colonization rates were uniform across all ethnic groups, implying the importance of environmental factors in H. pylori transmission in modern cities, independent of ethnicity.
Keywords: children, epidemiology, H. pylori, transmission
INTRODUCTION
The gastric bacterium Helicobacter pylori (H. pylori) colonizes more than half of the human population. It usually induces the influx of inflammatory cells in the stomach wall, which is a major risk factor for peptic ulcer disease and gastric cancer [1, 2], and also is associated with diminished risk of oesophageal reflux and childhood-onset asthma [3, 4, 5], and possibly more resistance to infectious diseases [6, 7]. H. pylori colonization is usually acquired during early childhood, and in most cases persists unless eliminated by antibiotic treatment [8]. A recent study reported that the risk of H. pylori colonization was influenced by host genetics [9].
The prevalence of colonization differs between children and adults [10]. Several cross-sectional surveys in Western countries have shown that H. pylori prevalence increases with age [11, 12]. Since acquisition during adulthood is rare [13, 14], the higher prevalence in the elderly rather reflects a birth cohort effect with higher rates of childhood exposure to the organism in the past [1]. The current lower levels of exposure to H. pylori and consequent lower prevalence in children are believed to be due to improved hygiene, and active elimination by antibiotics, together contributing to declining transmission risk [1, 15].
However, a recent study in Dutch children reported similarity in the H. pylori prevalence in two subsequent birth cohorts [16], possibly indicating that determinants previously responsible for declining colonization in the past now have stabilized. One factor contributing to this trend is the altered composition of western populations; during recent decades, the populations of western cities have become multi-ethnic as a result of immigration, often from countries where H. pylori remains endemic. Recently, we reported large differences in colonization rates among pregnant women of different ethnic origins living in Rotterdam, the Netherlands [17], but whether these differences are reflected in their offspring was not determined. Analysis of H. pylori transmission and risk factors would allow better prediction of the future incidence of H. pylori-associated illnesses.
In this population-based prospective cohort study, we aimed to measure H. pylori status, as well as risk factors for colonization and transmission, in children living in a multi-ethnic Western urban population, and in relation to colonization of their mothers. Unexpectedly, we found a relatively uniform intergenerational decrease in H. pylori prevalence in all nine ethnic groups studies. We explore the factors associated with this broad change.
METHODS
Design and setting
This study was embedded in the Generation R Study, a population-based prospective cohort study from fetal life onwards. All participants live in the multi-ethnic Rotterdam, the second largest city in the Netherlands. The children were born between April 2002 and January 2006. The background, design, and aims of the Generation R study have been reported in detail [18]. In total, 8,305 children and their parents participated in the postnatal phase of the study (from birth onwards) (Figure 1). From this initial population, 6,690 children visited the research centre at the age of 6 years. During these visits, blood samples were collected from 4,593 (69%) children (see Table S1 comparing the children, with and without H. pylori data). Data on age, ethnicity, breastfeeding, day-care attendance, antibiotic use, and socioeconomic status of the mother were collected using questionnaires. The Generation R Study was approved by the Medical Ethical Committee of the Erasmus University Medical Centre, and parents of the children gave written informed consent.
Covariates
The Generation R cohort comprises a wide range of ethnic groups, reflecting Rotterdam’s urban population as a typical Western city; the largest ethnic groups are of Dutch, Surinamese, Turkish, Moroccan, Dutch-Antilles, and Cape Verdean descent. Ethnicity is determined by country of birth of the child and its parents. According to the definition of Statistics Netherlands, a child was considered of non-Dutch ethnic origin if one of its parents was born abroad [19]. If both parents were born in two different countries outside the Netherlands, the country of the mother prevailed. Children of Dutch origin are considered as the native population. Participants with an ethnic background other than mentioned above are grouped together as European, Asian, African, or ‘rest of the world’, which included Central and South America (n = 95), Indonesia (n = 32), North America (n = 21), Asia, western (i.e. western ethnicity but parents had lived in Asia, n = 21), and Oceania (n = 9). Data on type of delivery was obtained by review of midwife and hospital registries. Exclusiveness of breastfeeding was categorized into three breastfeeding groups: never, non-exclusive breastfeeding until 4 months, or exclusive breastfeeding until 4 months. Data on day-care attendance was based on each child’s first year of life. Maternal parity served as a proxy for the presence of older siblings in the family, which was categorized as either none or at least one older sibling. Use of antibiotics was assessed by questionnaire at the ages of 12, 24, 36, 48, and 72 months. Parents were asked whether their child had received antibiotics (for example, penicillin) during the past year. Based on these questions, we calculated the cumulative number of courses in the first 6 years of life and computed a categorical variable with 3 groups: never any course of antibiotics, 1–2 courses, and 3 or more courses. Data on antibiotic use were not validated by physician prescriptions or pharmacy records. The socioeconomic status of the children was defined according to the educational level of their mother on the basis of her highest level of completed education. The highest educational level was defined as completion of university or higher vocational training. Mothers were categorized as having a middle-low level of education if they had completed intermediate vocational training, or had completed education below that level.
Serological determinants
Serum samples from 4,467 children, obtained at the mean age of 6.2 (±0.4) years (range 4–8 years), were available for analysis (Figure 1). The main reasons for missing blood samples were non-consent of the parents and technical or logistic failure. Procedures for the collection and storage of serum samples have been described [20]. Samples were examined for H. pylori IgG antibody levels by enzyme-linked immunosorbent assay (ELISA), using whole cell antigens [21]. A separate ELISA was performed to determine serum IgG antibodies against a specific recombinant truncated cytotoxin-associated gene A (CagA) protein, as described [22]. Both ELISAs have been validated in children [23], and have been previously used in Dutch children [16]. All samples were examined in duplicate; for each, the optical density ratio (ODR) was calculated by dividing the optical density (OD) by the mean OD of the positive controls. H. pylori-positive samples were those having either an ODR ≥ 1 or a CagA-positive test result. The cut-off value for CagA positivity was ODR value ≥ 0.35 [22]. Only 13 children were found to be CagA-positive but H. pylori-negative (0.3% of H. pylori-negative children). Based on prior studies, these subjects were considered as truly H. pylori positive [24]. Data on the maternal H. pylori status was available for 3,185 children (71%). H. pylori antibody distributions for both children and their mothers are shown in Figure S1. Details regarding H. pylori colonization in the total cohort of mothers have been described [17].
Statistical analysis
Chi-square tests (categorical variables) and t-tests (continuous variables) were used to compare different variables with H. pylori status. Univariate analyses were performed to assess determinants associated with H. pylori presence. To study the individual effect of each potential determinant, each was tested separately, followed by a multivariate analysis corrected for all others. For all covariates, the percent of missing values within the population for analysis was lower than 10%, except for caesarean section, breastfeeding, day-care attendance, and antibiotic use. Missing data in the covariates (except maternal H. pylori status) were imputed with multiple imputations using chained equations, by which the most likely value for a missing response is selected [25]. Ten new datasets were created by imputation based on all covariates and outcomes in the model. Data from each separate imputation was analysed, after which results were combined. Except for breastfeeding and antibiotic exposure, no major differences in the direction or magnitude of the effect estimates were observed between analyses with imputed missing data and complete cases only. Only the results based on the pooled imputed datasets are presented in this manuscript. To identify the potential modifying effect of determinants included in the multivariate analyses, we evaluated statistical interaction by adding to the multivariate model the product term of an independent variable and subgroup (independent variable x subgroup) as covariate. The interaction was tested between ethnicity and day-care, ethnicity and gender, and ethnicity and maternal education level. We calculated the population attributable fraction (PAF) for H. pylori disappearance in children with an H. pylori-positive mother, using adjusted odds ratios estimated from logistic regression models [26]. All measures of associations are presented as Odds Ratios (OR) with their 95% Confidence Intervals (CI). Statistical analyses were performed using IBM SPSS Statistics 21.0 for Windows (SPSS IBM, Armonk, New York, USA).
RESULT
H. pylori prevalence
The serum of 4,467 children was analysed (Figure 1). Table 1 summarizes both the observed population characteristics as well as the imputed data, stratified by H. pylori status. Overall, 438 (10%; 95% CI 8.9–10.7%) children were H. pylori-positive, of whom 142 (32%; 95% CI 28.0–36.8%) also were CagA-positive. In all children of non-Dutch ethnicity, the colonization rate was significantly higher than that in children of Dutch ethnicity (Figure 2A). Higher H. pylori colonization rates were observed in children of either Dutch or non-Dutch ethnicity of older age (Figure S2). The proportion of CagA-positivity amongst H. pylori-positive children varied widely between ethnic groups (Figure 2B). The lowest proportions were found in children with Dutch or other European ethnicities (16% and 14%, respectively).
Table 1.
Child characteristics | Observed | Imputed |
H. pylori− n = 4,029 |
H. pylori+ n = 438 |
Univariate OR (95% CI) |
---|---|---|---|---|---|
Female sex (%) | 2,164 (48.4) | 2,164 (48.4) | 1,916 (47.6) | 248 (56.6) | 1.44 (1.18–1.76)* |
Mean age sera taken, years (SD) | 6.2 (0.4) | 6.2 (0.4) | 6.2 (0.01) | 6.4 (0.03) | 1.66 (1.39–1.98)* |
Ethnicity (%) | |||||
Dutch | 2,505 (56.1) | 2,516 (56.3) | 2 ,374 (58.9) | 142 (32.4) | 1.0 |
Surinamese | 317 (7.1) | 318 (7.1) | 285 (7.1) | 33 (7.5) | 1.95 (1.29–2.94)* |
Turkish | 311 (7.0) | 324 (7.3) | 279 (6.9) | 45 (10.3) | 2.64 (1.71–4.07)* |
Moroccan | 256 (5.7) | 261 (5.8) | 192 (4.8) | 69 (15.8) | 6.01 (4.27–8.44)* |
Dutch Antilles | 141 (3.2) | 142 (3.2) | 120 (3.0) | 21 (4.8) | 2.96 (1.78–4.92)* |
Cape Verdean | 129 (2.9) | 151 (3.4) | 115 (2.9) | 36 (8.2) | 5.20 (3.31–8.16)* |
Other: | |||||
European | 331 (7.4) | 332 (7.4) | 298 (7.4) | 34 (7.8) | 1.92 (1.28–2.89)* |
Asian | 111 (2.5) | 112 (2.5) | 95 (2.4) | 17 (3.9) | 3.01 (1.73–5.23)* |
African | 100 (2.2) | 106 (2.4) | 91 (2.3) | 15 (3.4) | 2.63 (1.31–5.30)* |
Rest of the worlda | 148 (3.3) | 206 (4.6) | 180 (4.5) | 26 (5.9) | 2.11 (0.74–6.00) |
Data missing | 118 (2.6) | ||||
Caesarean Section (%) | |||||
No | 3,358 (75.2) | 3,878 (86.8) | 3,483 (86.4) | 395 (90.2) | 1.0 |
Yes | 496 (11.1) | 589 (13.2) | 546 (13.6) | 43 (9.8) | 0.70 (0.47–1.02) |
Data missing | 613 (13.7) | ||||
Breastfeedingb (%) | |||||
Never | 264 (5.9) | 388 (8.7) | 356 (8.8) | 32 (7.3) | 1.0 |
Partial | 1,857 (41.6) | 2,925 (65.5) | 2,623 (65.1) | 302 (68.9) | 1.34 (0.73–2.43) |
Exclusive | 723 (16.2) | 1,154 (25.8) | 1,050 (26.1) | 104 (23.7) | 1.14 (0.64–2.03) |
Data missing | 1,623 (36.3) | ||||
Day-care attendancec (%) | |||||
No | 1,019 (22.8) | 2,110 (47.2) | 1,849 (45.9) | 261 (59.6) | 1.0 |
Yes | 1,548 (34.7) | 2,357 (52.8) | 2,180 (54.1) | 177 (40.4) | 0.57 (0.44–0.75)* |
Data missing | 1,900 (42.5) | ||||
Number of older siblings (%) | |||||
0 | 2,370 (53.1) | 2,425 (54.3) | 2,219 (55.1) | 206 (47.0) | 1.0 |
≥1 | 1,936 (43.3) | 2,042 (45.7) | 1,810 (44.9) | 232 (53.0) | 1.38 (1.13–1.69)* |
Data missing | 161 (3.6) | ||||
Antibiotic exposure (%) | |||||
6–11 months | |||||
No | 1,800 (40.3) | 2,728 (61.1) | 2,467 (61.2) | 261 (59.6) | 1.0 |
1–2 courses | 879 (19.7) | 1,390 (31.1) | 1,263 (31.3) | 127 (29.0) | 0.95 (0.68–1.32) |
≥3 courses | 105 (2.4) | 350 (7.8) | 300 (7.4) | 50 (11.4) | 1.51 (0.82–2.77) |
Data missing | 1,683 (37.7) | ||||
12–23 months | |||||
No | 1,638 (36.7) | 2,337 (52.3) | 2,114 (52.5) | 223 (50.9) | 1.0 |
1–2 courses | 1,111 (24.9) | 1,702 (38.1) | 1,541 (38.2) | 161 (36.8) | 0.99 (0.74–1.32) |
≥3 courses | 195 (4.4) | 428 (9.6) | 374 (9.3) | 54 (12.3) | 1.35 (0.86–2.11) |
Data missing | 1,523 (34.1) | ||||
24–35 months | |||||
No | 1,817 (40.7) | 2,642 (59.1) | 2,404 (59.7) | 238 (54.3) | 1.0 |
1–2 courses | 886 (19.8) | 1,521 (34.0) | 1,364 (33.9) | 156 (35.6) | 1.16 (0.87–1.56) |
≥3 courses | 105 (2.4) | 305 (6.8) | 260 (6.5) | 44 (10.0) | 1.71 (0.99–2.95) |
Data missing | 1,659 (37.1) | ||||
36–47 months | |||||
No | 1,977 (44.3) | 2,820 (63.1) | 2,566 (63.7) | 254 (58.0) | 1.0 |
1–2 courses | 763 (17.1) | 1,253 (28.1) | 1,131 (28.1) | 122 (27.9) | 1.09 (0.79–1.50) |
≥3 courses | 75 (1.7) | 395 (8.8) | 333 (8.3) | 62 (14.2) | 1.80 (0.91–3.57) |
Data missing | 1,652 (37.0) | ||||
60–71 months | |||||
No | 2,956 (66.2) | 3,303 (73.9) | 2,999 (74.4) | 304 (69.4) | 1.0 |
1–2 courses | 760 (17.0) | 921 (20.6) | 829 (20.6) | 93 (21.2) | 1.10 (0.83–1.46) |
≥3 courses | 73 (1.6) | 243 (5.4) | 202 (5.0) | 41 (9.4) | 1.95 (1.03–3.67)* |
Data missing | 678 (15.2) | ||||
Cumulative antibiotic use | |||||
Never | 529 (11.8) | 940 (21.0) | 861 (21.4) | 78 (17.8) | 1.0 |
1–2 courses | 1,162 (26.0) | 1,715 (38.4) | 1,564 (38.8) | 151 (34.5) | 1.07 (0.73–1.56) |
≥3 courses | 828 (18.5) | 1,813 (40.6) | 1,604 (39.8) | 209 (47.7) | 1.44 (1.01–2.06)* |
Data missing | 1,948 (43.6) | ||||
Maternal characteristics | |||||
| |||||
Maternal education level (%) | |||||
Primary, or secondary | 2101 (47.0) | 2,438 (54.6) | 2,127 (52.8) | 311 (71.0) | 2.20 (1.76–2.75)* |
Higher | 1955 (43.8) | 2,029 (45.4) | 1,902 (47.2) | 127 (29.0) | 1.0 |
Data missing | 411 (9.2) | ||||
Mean age sera taken, years (SD) | 30.5 (5.0) | 30.5 (5.0) | 31.3 (4.6) | 29.4 (5.4) | 0.92 (0.91–0.94)* |
Values are means (and standard deviation), absolute numbers (and percentages) or odds ratio (and 95% confidence interval). Missing data on maternal H. pylori and CagA status were not imputed.
Includes subjects from Central and South America (n = 95), Indonesia (n = 32), North America (n = 21), Asia, western (n = 21), and Oceania (n = 9).
Data until 4 months of life
Data completed from the first year of life. * p < 0.05
Risk factors for H. pylori colonization
In univariate analyses, a child’s H. pylori positivity was associated with an H. pylori-positive mother (OR 3.22; 95% CI 2.52–4.12), and non-Dutch ethnicity (OR 2.99; 95% CI 2.32–3.86). Female gender, age, breastfeeding, presence of older siblings, antibiotic exposure, and lower maternal educational level also were positively associated with H. pylori status, whereas day-care attendance was negatively associated (Table 1).
Using multivariate analysis (Figure 3), we identified the following independent risk factors for H. pylori positivity in a child: maternal H. pylori positivity [(CagA-positive mother OR 2.25; 95% CI 1.61–3.16) (CagA-negative mother OR 2.05; 95% CI 1.53–2.74)], non-Dutch ethnicity (OR 2.04; 95% CI 1.53–2.72), female gender (OR 1.47; 95% CI 1.20–1.81), and lower maternal education level (OR 1.37; 95% CI 1.06–1.78). A separate multivariate analysis to examine risk for CagA-positivity amongst all H. pylori-positive children revealed independent associations with lower educational level of mother (OR 2.65; 95% CI 1.33–5.28), and non-Dutch ethnicity (OR 2.48; 95% CI 1.27–4.85) (Table S3). Compared with males, we found female gender independently associated with never having had exposure to antibiotics (OR 1.30; 95% CI 1.07–1.60) and lower educational level of the mother (OR 1.17; 95% CI 1.01–1.35) (Table S4). Caesarian section was not significantly associated with H. pylori colonization. Comparison of C–section with vaginal birth revealed independent associations with no breastfeeding (OR 1.98; 95% CI 1.31–2.98), nulliparity (OR 1.81; 95% CI 1.48–2.20), and day-care attendance (OR 1.34; 95% CI 1.01–1.77) (Table S5).
A stratified analysis by ethnicity was performed (Dutch vs. non-Dutch), based on the significantly lower H. pylori colonization rate in children of Dutch ethnicity compared to all other subjects (Figure 3, and Table S6 for comparison of European vs. non-European). Differences in the odds ratios for H. pylori colonization were observed for gender, educational level of mother, and day-care attendance. There was no evidence for effect modification by ethnicity for the associations of gender, educational level of mother, and day-care attendance with H. pylori colonization (p-value for interactions >0.05).
Comparison of H. pylori colonization in mothers and their children
Data on the H. pylori status of mother was available for 3,185 (71%) children (Table 2). The H. pylori positivity rate in mothers (mean age of 30.5 ±5.0 years) was 42%. An H. pylori-positive mother was associated with an H. pylori-positive child (OR 3.22; 95% CI 2.52–4.12). Of the 1,328 children with an H. pylori-positive mother, 211 (15.9%) were H. pylori-positive, compared to 103 (5.5%) of the 1,857 children with an H. pylori-negative mother. As a result, 33% (n=103) of all H. pylori-positive children had a mother who tested H. pylori-negative. The median antibody titer in these children was significantly lower (1.43; 2.5–97.5th percentile 0.74–6.68) than in children with an H. pylori-positive mother (2.11; 2.5–97.5th percentile 0.56–15.68). In children of non-Dutch ethnicity, the proportion H. pylori-positive children with an H. pylori-negative mother was 22% compared to 55% of children with Dutch ethnicity (p<0.001). Table 2 shows the associations between mother and child’s H. pylori colonization rates by strain type. Children born from H. pylori-negative mothers were significantly less likely to be colonized at age 6 with either a CagA-negative or CagA-positive H. pylori strain than children born from H. pylori-positive mothers. Conversely, children born from H. pylori+CagA− mothers were more likely to be colonized with the same strain, and for children with H. pylori+CagA+ mothers, the OR for carrying the same strain was 6.74 (95% CI 4.52–10.05).
Table 2.
Child’s H. pylori status | ||||
---|---|---|---|---|
Hp− (n= 2,871) |
Hp+CagA− (n=210) |
Hp+CagA+ (n=104) |
||
Mother’s H. pylori status | Hp− [n= 1,857 (%)] OR (95% CI) |
1,754 (94.5) Reference |
86 (4.6) 0.44 (0.33–0.59)* |
17 (0.9) 0.12 (0.07–0.21)* |
Hp+CagA− [n= 873 (%)] OR (95% CI) |
746 (85.5) Reference |
92 (10.5) 2.22 (1.67–2.95)* |
35 (4.0) 1.45 (0.95–2.19) |
|
Hp+CagA+ [n= 455 (%)] OR (95% CI) |
371 (81.5) Reference |
32 (7.0) 1.21 (0.82–1.79) |
52 (11.5) 6.74 (4.52–10.05)* |
Values shown are absolute numbers and their percentages in relation to maternal H. pylori status. The odds ratios and 95% confidence intervals represent the association between the reference group (children without H. pylori) and the other groups.
p <0.05
Overall, the H. pylori prevalence decreased 76% comparing mothers and their children. A significant decline in H. pylori prevalence was observed across all nine ethnic groups studied (Figure 4). This decline was consistent for both H. pylori+CagA− and H. pylori+CagA+-strains (Figure 4A and 4B). The overall decline rate for males (−80%) was slightly higher than for females (−72%), which was consistent across all ethnic groups. Multivariate analysis of the loss of H. pylori in children with an H. pylori-positive mother (n = 1,328) revealed male gender (OR 1.64; 95% CI 1.21–2.23), higher maternal education level (OR 1.78; 95% CI 1.15–2.76), and no older siblings (OR 1.37; 95% CI 1.01–1.88) independently associated with an H. pylori-negative child (Table S7). The proportion of the H. pylori decline attributable to male gender (21%), having no older siblings (14%), and higher maternal education level (14%), were all significant (Figure 5).
DISCUSSION
In this multi-ethnic population-based cohort, we found highly variable H. pylori colonization rates in six-year old children, with both the prevalence of H. pylori and the proportion of CagA-positive strains higher in children of non-Dutch ethnicity. Independent of ethnic background, maternal H. pylori colonization was the strongest risk factor for H. pylori-positivity in their offspring. Our study design made it possible to compare H. pylori colonization in children directly with that of their mothers, showing essentially identical intergenerational reductions for both H. pylori+CagA− and H. pylori+CagA+-strains.
The overall colonization rate of 10% differs from that found in a previous study performed in the Netherlands [16]. This recent study of 545 Dutch children between 7 and 9 years old, which used the same ELISA as did we, found an H. pylori positivity rate of 9% (95% CI 6.6–11.4%) [16], a prevalence slightly higher than that measured in children of Dutch ethnicity in our study (6%). This difference may reflect a continuing decline in colonization, or may be due to the different study designs, or the younger age of children in our study. The latter may contribute, as our data suggest continuing acquisition of H. pylori at least until the age of 7, consistent in children of Dutch and non-Dutch ethnicities. The higher H. pylori prevalence in children of non-Dutch ethnicity confirms findings of other studies [27, 28]. When comparing previous studies, the exact colonization rate of particular ethnic groups may differ due to differences in age or selected populations, but nevertheless it is clear that subjects of non-Western ethnicity comprise risk groups for H. pylori colonization within multi-ethnic populations of Western cities.
The intergenerational decrease was of the same magnitude among all different ethnic groups, resulting in the same birth cohort effect in all groups. The decline in children with an H. pylori-positive mother can be partially attributed to male gender, lack of older siblings, and higher educational level of mother. The lack of older siblings may reduce horizontal transmission of H. pylori within a family. Others found the number of siblings within a family rather than birth order independently associated with H. pylori colonization [29]. Nevertheless, our findings imply that environmental factors and living conditions of the country in which a child is raised have a major impact on transmission, irrespective of ethnicity. The consistent decline across all ethnic groups support the hypothesis that in contemporary Dutch society, and probably elsewhere as well, there are highly prevalent factors that interfere with the early life acquisition and-or maintenance of H. pylori. Besides the involvement of socio-economic status, family size, and other living conditions, possible candidates include the widespread use of antibiotics, particularly in young children. The effect of antibiotic monotherapy on H. pylori status is limited [30], but repeated antibiotic exposure may eventually result in eradication. Another possibility could be the run-off of antibiotics from farms where antibiotic-intensive husbandry is being practiced, which may contaminate surface and drinking water [31]. In contrast to the use of antibiotics in humans, the Dutch consumption of antibiotics per animal exceeds the consumption of all European countries, and despite the prohibition of antibiotics as growth promoters, the use remained stable [32]. There is no strong evidence of contaminated drinking water; in a recent screening of superficial groundwater used for the production of drinking water, none of the veterinary pharmaceuticals have been observed, and no concentrations of related compounds were observed above the threshold of toxicological concern [32]. In a separate survey of human pharmaceuticals, clindamycin (5 ng/L) and erythromycin (10 ng/L) residuals were found in surface water, but not in produced drinking water [32]. Regardless of its cause, the clinical consequences of this rapid disappearance may have opposite effects: a fall in prevalence of the later life expression of gastric and duodenal ulcer disease, and gastric carcinoma [33, 34], but a rise in earlier life-expressed atopy, asthma, and reflux-related disorders [11], since epidemiological studies have shown inverse associations of these disorders with H. pylori colonization [3, 5].
The association of specific H. pylori types in mother and child provides further evidence supporting a role for maternal inheritance in early life transmission, shown in molecular typing studies [35]. A recent German study that included the H. pylori status of parents and siblings in a multivariate model, showed that only maternal infection was associated with H. pylori positivity in the children [28]. Despite this important maternal role, we found that one third of all positive children had an apparently H. pylori-negative mother. This proportion was even higher (>50%) in children of Dutch ethnicity, implying the involvement of other transmission sources, such as fathers and siblings [35]. An alternative hypothesis is that some maternal H. pylori colonisations were missed, due to lack of complete sensitivity of the assay, or post-natal acquisition of the organism.
Especially in a multi-ethnic population, children may become colonized with H. pylori by acquiring the bacterium from persons of ethnicities with higher H. pylori prevalence, e.g., in day-care facilities. However, a recently published meta-analysis found no significant effect of day-care attendance on H. pylori colonization (summary OR 1.12; 95% CI 0.82–1.52) [36]. Nevertheless, a Portuguese study of 1,047 children reported increasing H. pylori prevalence with cumulative attendance in day-care centres [37]. Our stratified analysis of ethnicity revealed opposite trends for the relation between H. pylori colonization and day-care attendance. Such observations suggest that child-to-child transmission in a day-care setting may be more likely for children of Dutch ethnicity where children of non-Dutch ethnicity could serve as transmission sources.
Remarkably, female gender was found to be associated with H. pylori colonization; a possible explanation may be the higher antibiotic exposures we observed in males. This may also explain the higher rate of decline we found in boys compared to girls. The positive associations of Caesarian section with nulliparity, and no breastfeeding, confirm prior observations [38], but that mothers who underwent Caesarian section were more likely to use day-care suggests that mode of delivery correlates with other lifestyle aspects.
An important strength of this study is that we had a large multi-ethnic study population drawn from the general population of Rotterdam; since immigration is common in many western countries, our findings may be more broadly applicable. An additional strength is the use of maternal data on H. pylori colonization, which provides insight into mother-to-child transmission.
This study has some limitations, including missing data for several characteristics and potential risk factors for colonization, which may have biased the outcome. However, we performed the final analyses after a multiple imputation procedure, considered useful to deal with missing data, as it requires the fewest assumptions and reduces potential bias when missing data are not random [25]. A second limitation was lack of data on H. pylori colonization in fathers and siblings, precluding examination of their potential roles in transmission. A third limitation is that data on antibiotic exposures were not validated by pharmacy records, nor was information available on specific types. Finally, although both ELISAs have been validated in adults and children, including Dutch adults, and have been used in previous studies in Dutch children [16], they have not been separately validated in Dutch children.
In conclusion, we found relatively high H. pylori colonization rates in children of non-Dutch ethnicity who were born and raised in a western city. Regardless of ethnicity, maternal H. pylori type was an important predictor for a child’s H. pylori type. The high and consistent intergenerational decline in H. pylori prevalence irrespective of ethnicity and sex points toward very common exposures fuelling this phenomenon.
Supplementary Material
SUMMARY BOX.
- What is already known about this subject?
- H. pylori prevalence of children living in Western countries is low.
- Maternal H. pylori status is an important transmission source for H. pylori colonization in their children.
- Migrant communities in Western populations constitute risk groups for H. pylori colonization.
- What are the new findings?
- A high intergenerational decline in H. pylori prevalence was found, comparing mothers with their children, with nearly identical rates (76% and 77%) for Hp+CagA− and Hp+CagA+/strains, respectively.
- The intergenerational drop in H. pylori prevalence was uniform in nine separate ethnicities.
- Risk factors for H. pylori positivity are mostly the same among diverse ethnic groups.
- Our data suggest a continuing acquisition of H. pylori at least to age 7.
- How might it impact on clinical practice in the foreseeable future?
- The maternal-child linkage is to some degree predictive of H. pylori positivity in a child, which affects risk of subsequent diseases.
Acknowledgments
The Generation R Study is conducted by the Erasmus Medical Centre in close collaboration with the School of Law and Faculty of Social Sciences of the Erasmus University Rotterdam; the Municipal Health Service Rotterdam area, Rotterdam; the Rotterdam Homecare Foundation, Rotterdam; and the Stichting Trombosedienst and Artsenlaboratorium Rijnmond (STAR), Rotterdam. The authors gratefully acknowledge the contribution of participating parents, children, general practitioners, hospitals, midwives, and pharmacies in Rotterdam.
FUNDING
Supported in part by R01DK090989 from the National Institutes of Health, the Diane Belfer Program of Human Microbial Ecology, and by the Knapp Family Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Footnotes
COMPETING INTERESTS
None of the authors has any conflict of interest.
References
- 1.Kusters JG, van Vliet AH, Kuipers EJ. Pathogenesis of Helicobacter pylori infection. Clin Microbiol Rev. 2006;19:449–90. doi: 10.1128/CMR.00054-05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Blaser MJ, Nomura A, Lee J, et al. Early-life family structure and microbially induced cancer risk. PLoS Med. 2007;4:e7. doi: 10.1371/journal.pmed.0040007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Loffeld RJ, Werdmuller BF, Kuster JG, et al. Colonization with CagA-positive Helicobacter pylori strains inversely associated with reflux esophagitis and Barrett’s esophagus. Digestion. 2000;62:95–9. doi: 10.1159/000007801. [DOI] [PubMed] [Google Scholar]
- 4.Blaser MJ, Chen Y, Reibman J. Does Helicobacter pylori protect against asthma and allergy? Gut. 2008;57:561–7. doi: 10.1136/gut.2007.133462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chen Y, Blaser MJ. Helicobacter pylori colonization is inversely associated with childhood asthma. J Infect Dis. 2008;198:553–60. doi: 10.1086/590158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rothenbacher D, Blaser MJ, Bode G, et al. Inverse relationship between gastric colonization of Helicobacter pylori and diarrheal illnesses in children: results of a population-based cross-sectional study. J Infect Dis. 2000;182:1446–9. doi: 10.1086/315887. [DOI] [PubMed] [Google Scholar]
- 7.Perry S, de Jong BC, Solnick JV, et al. Infection with Helicobacter pylori is associated with protection against tuberculosis. PLoS One. 2010;5:e8804. doi: 10.1371/journal.pone.0008804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Atherton JC, Blaser MJ. Coadaptation of Helicobacter pylori and humans: ancient history, modern implications. J Clin Invest. 2009;119:2475–87. doi: 10.1172/JCI38605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Mayerle J, den Hoed CM, Schurmann C, et al. Identification of genetic loci associated with Helicobacter pylori serologic status. JAMA. 2013;309:1912–20. doi: 10.1001/jama.2013.4350. [DOI] [PubMed] [Google Scholar]
- 10.Koletzko S, Jones NL, Goodman KJ, et al. Evidence-based guidelines from ESPGHAN and NASPGHAN for Helicobacter pylori infection in children. J Pediatr Gastroenterol Nutr. 2011;53:230–43. doi: 10.1097/MPG.0b013e3182227e90. [DOI] [PubMed] [Google Scholar]
- 11.Blaser MJ, Falkow S. What are the consequences of the disappearing human microbiota? Nat Rev Microbiol. 2009;7:887–94. doi: 10.1038/nrmicro2245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Murray LJ, McCrum EE, Evans AE, et al. Epidemiology of Helicobacter pylori infection among 4742 randomly selected subjects from Northern Ireland. Int J Epidemiol. 1997;26:880–7. doi: 10.1093/ije/26.4.880. [DOI] [PubMed] [Google Scholar]
- 13.Kuipers EJ, Pena AS, van Kamp G, et al. Seroconversion for Helicobacter pylori. Lancet. 1993;342:328–31. doi: 10.1016/0140-6736(93)91473-y. [DOI] [PubMed] [Google Scholar]
- 14.Malaty HM, El-Kasabany A, Graham DY, et al. Age at acquisition of Helicobacter pylori infection: a follow-up study from infancy to adulthood. Lancet. 2002;359:931–5. doi: 10.1016/S0140-6736(02)08025-X. [DOI] [PubMed] [Google Scholar]
- 15.Blaser MJ, Atherton JC. Helicobacter pylori persistence: biology and disease. J Clin Invest. 2004;113:321–33. doi: 10.1172/JCI20925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.den Hoed CM, Vila AJ, Holster IL, et al. Helicobacter pylori and the birth cohort effect: evidence for stabilized colonization rates in childhood. Helicobacter. 2011;16:405–9. doi: 10.1111/j.1523-5378.2011.00854.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.den Hollander WJ, Holster IL, den Hoed CM, et al. Ethnicity is a strong predictor for Helicobacter pylori infection in young women in a multi-ethnic European city. J Gastroenterol Hepatol. 2013;28:1705–11. doi: 10.1111/jgh.12315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jaddoe VW, van Duijn CM, Franco OH, et al. The Generation R Study: design and cohort update 2012. Eur J Epidemiol. 2012;27:739–56. doi: 10.1007/s10654-012-9735-1. [DOI] [PubMed] [Google Scholar]
- 19.Netherlands S. Allochtonen in Nederland (in Dutch) Voorburg-Heerlen; 2004. [Google Scholar]
- 20.Jaddoe VW, Bakker R, van Duijn CM, et al. The Generation R Study Biobank: a resource for epidemiological studies in children and their parents. Eur J Epidemiol. 2007;22:917–23. doi: 10.1007/s10654-007-9209-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Perez-Perez GI, Dworkin BM, Chodos JE, et al. Campylobacter pylori antibodies in humans. Ann Intern Med. 1988;109:11–7. doi: 10.7326/0003-4819-109-1-11. [DOI] [PubMed] [Google Scholar]
- 22.Blaser MJ, Perez-Perez GI, Kleanthous H, et al. Infection with Helicobacter pylori strains possessing cagA is associated with an increased risk of developing adenocarcinoma of the stomach. Cancer Res. 1995;55:2111–5. [PubMed] [Google Scholar]
- 23.Drumm B, Perez-Perez GI, Blaser MJ, et al. Intrafamilial clustering of Helicobacter pylori infection. N Engl J Med. 1990;322:359–63. doi: 10.1056/NEJM199002083220603. [DOI] [PubMed] [Google Scholar]
- 24.Romero-Gallo J, Perez-Perez GI, Novick RP, et al. Responses of endoscopy patients in Ladakh, India, to Helicobacter pylori whole-cell and Cag A antigens. Clin Diagn Lab Immunol. 2002;9:1313–7. doi: 10.1128/CDLI.9.6.1313-1317.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338:b2393. doi: 10.1136/bmj.b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions. Am J Public Health. 1998;88:15–9. doi: 10.2105/ajph.88.1.15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Miendje Deyi VY, Vanderpas J, Bontems P, et al. Marching cohort of Helicobacter pylori infection over two decades (1988–2007): combined effects of secular trend and population migration. Epidemiol Infect. 2011;139:572–80. doi: 10.1017/S095026881000110X. [DOI] [PubMed] [Google Scholar]
- 28.Weyermann M, Rothenbacher D, Brenner H. Acquisition of Helicobacter pylori infection in early childhood: independent contributions of infected mothers, fathers, and siblings. Am J Gastroenterol. 2009;104:182–9. doi: 10.1038/ajg.2008.61. [DOI] [PubMed] [Google Scholar]
- 29.Ford AC, Forman D, Bailey AG, et al. Effect of sibling number in the household and birth order on prevalence of Helicobacter pylori: a cross-sectional study. Int J Epidemiol. 2007;36:1327–33. doi: 10.1093/ije/dym201. [DOI] [PubMed] [Google Scholar]
- 30.Broussard CS, Goodman KJ, Phillips CV, et al. Antibiotics taken for other illnesses and spontaneous clearance of Helicobacter pylori infection in children. Pharmacoepidemiol Drug Saf. 2009;18:722–9. doi: 10.1002/pds.1773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Blaser MJ. Who are we? Indigenous microbes and the ecology of human diseases. EMBO Rep. 2006;7:956–60. doi: 10.1038/sj.embor.7400812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Hofman-Caris CHM, de Jongh CM, Wols BA, et al. Dealing with pharmaceuticals in drinking water production. 025. KWR; BTO; 2012. p. 2013. [Google Scholar]
- 33.Sung JJ, Kuipers EJ, El-Serag HB. Systematic review: the global incidence and prevalence of peptic ulcer disease. Aliment Pharmacol Ther. 2009;29:938–46. doi: 10.1111/j.1365-2036.2009.03960.x. [DOI] [PubMed] [Google Scholar]
- 34.Forman D, Burley VJ. Gastric cancer: global pattern of the disease and an overview of environmental risk factors. Best Pract Res Clin Gastroenterol. 2006;20:633–49. doi: 10.1016/j.bpg.2006.04.008. [DOI] [PubMed] [Google Scholar]
- 35.Raymond J, Thiberg JM, Chevalier C, et al. Genetic and transmission analysis of Helicobacter pylori strains within a family. Emerg Infect Dis. 2004;10:1816–21. doi: 10.3201/eid1010.040042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Bastos J, Carreira H, La Vecchia C, et al. Childcare attendance and Helicobacter pylori infection: systematic review and meta-analysis. Eur J Cancer Prev. 2013;22:311–9. doi: 10.1097/CEJ.0b013e32835b69aa. [DOI] [PubMed] [Google Scholar]
- 37.Lunet N, Peleteiro B, Bastos J, et al. Child day-care attendance and Helicobacter pylori infection in the Portuguese birth cohort Geracao XXI. Eur J Cancer Prev. 2013 doi: 10.1097/CEJ.0b013e328364742a. [DOI] [PubMed] [Google Scholar]
- 38.Prior E, Santhakumaran S, Gale C, et al. Breastfeeding after cesarean delivery: a systematic review and meta-analysis of world literature. Am J Clin Nutr. 2012;95:1113–35. doi: 10.3945/ajcn.111.030254. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.