Helicobacter pylori causes antral gastritis and is associated. Low Rates of Helicobacter pylori Reinfection in Children. Materials and Methods

Transcription

1 GASTROENTEROLOGY 1999;117: Low Rates of Helicobacter pylori Reinfection in Children MARION ROWLAND,* DEEPAK KUMAR,* LESLIE DALY, PAMELA O CONNOR, DAVID VAUGHAN, and BRENDAN DRUMM*, *Department of Paediatrics and Department of Public Health Medicine and Epidemiology, University College Dublin, and The Children s Research Centre, Our Lady s Hospital for Sick Children, Dublin, Ireland Background & Aims: Reinfection after treatment for Helicobacter pylori is uncommon in adults. It is more likely to occur in children because they acquire primary infection. The aim of this study was to determine whether children are likely to become reinfected with H. pylori and if there are any risk factors for reinfection. Methods: A prospective study of children who had documented evidence of successful treatment for H. pylori infection was performed. Sixty children were eligible for inclusion; results for 52 are presented. Children, parents, and siblings underwent [ 13 C]urea breath tests. Details of family size and socioeconomic status were documented. Cox logistic regression analysis was used to determine the risk factors for reinfection. Results: The duration of follow-up was patient-years (mean SD, months). Fortysix (88.5%) of the index children remained clear of infection, and 6 (11.5%) children were reinfected. The mean age of those who became reinfected was years compared with years for those who remained clear of infection (P ). Only 2 of 46 (4.3%) children older than 5 years of age were reinfected, although 80.8% had 1 infected parent and 65% of siblings were infected. Reinfection rate was 2.0% per person per year in children older than 5 years. Living with infected parents and siblings and low socioeconomic status were not risk factors for reinfection. In logistic regression analysis, age was the only risk factor for reinfection. Conclusions: Reinfection with H. pylori occurs rarely in children older than 5 years of age regardless of socioeconomic group or number of infected family members. These findings also indicate that it is not necessary to treat all family members to achieve long-term eradication of H. pylori. Helicobacter pylori causes antral gastritis and is associated with duodenal ulcer disease in adults 1 4 and children. 5 Eradication of the organism leads to long-term healing of duodenal ulcer disease. 6 9 The World Health Organization recently classified H. pylori as a group 1 carcinogen for gastric adenocarcinoma. 10 It is now accepted that H. pylori infection is almost always acquired in childhood and that primary infection in adults is rare Unless treatment is instituted, H. pylori infection persists throughout life in most individuals. 13 The increasing prevalence seen with increasing age is probably a birth cohort effect, 11,13 reflecting higher rates of acquisition during childhood earlier this century when adults lived in poorer socioeconomic conditions. Treatment of infected children should reduce the transmission of infection and ultimately reduce the incidence of peptic ulcer disease and gastric cancer. However, the effectiveness of such treatment in reducing H. pylori related morbidity is critically dependent on the rate of reinfection after treatment in children. Reinfection in adults is uncommon, varying from 0.64% 14 to 1.2% 15 per year. However, children are more likely to become reinfected after successful treatment of H. pylori because they are most at risk of acquiring primary infection. It is essential to determine the reinfection rate in children before any consensus on widespread treatment of H. pylori infection can be achieved. This knowledge will also help determine if there is a need to treat all family members of patients with duodenal ulcer disease and H. pylori. The aim of this study was to determine if children become reinfected with H. pylori after treatment and to evaluate possible risk factors predisposing to reinfection. Materials and Methods A prospective study of children treated successfully for H. pylori infection at Our Lady s Hospital for Sick Children, Dublin, Ireland, between February 1991 and July 1996 was performed. The indications for initial gastroduodenoscopy included abdominal pain, failure to thrive, vomiting, and hematemesis. H. pylori infection was diagnosed if children had a positive culture or if both histological examination and rapid urease test (Clo, Delta West, Australia) results were positive. Children were treated according to treatment protocols in place at the time of diagnosis of infection. Twenty-three children were treated with colloidal bismuth subcitrate, clarithromycin, and metronidazole for 1 week. 16 Twenty-one children received colloidal bismuth subcitrate and metronida- Abbreviation used in this paper: UBT, [ 13 C]urea breath test by the American Gastroenterological Association /99/$10.00

2 August 1999 HELICOBACTER PYLORI REINFECTION IN CHILDREN 337 zole for 2 weeks. 17 Five children were treated with colloidal bismuth subcitrate and amoxicillin for 2 weeks, and 2 children were treated with omeprazole and amoxicillin for 2 weeks. One child in whom two courses of treatment failed was successfully treated with colloidal bismuth subcitrate, tetracycline, and metronidazole for 4 weeks. All children were evaluated for H. pylori eradication 4 6 weeks after completion of treatment. Before July 1995, children (n 33) were considered successfully treated if culture, urease testing, and staining of histological sections all failed to identify the organism. Children who had one or more positive biopsy-based test results were not included in the study. After July 1995, children (n 22) were included if they had a negative [ 13 C]urea breath test (UBT) result at least 4 weeks after completion of treatment. Testing for H. pylori reinfection using the UBT began in July Before this date, no long-term assessment of reinfection could be ethically undertaken in children because the UBT was not validated in children. 18 Therefore, there is a variable length of follow-up for children in this study. The first breath test for reinfection took place at least 6 months after successful treatment, and breath tests were repeated annually. A researcher-administered questionnaire was used to document demographic details and symptoms, including abdominal pain, nausea, vomiting, and night awakening for each index child. All medications used within the previous month, including antibiotics, antacids, or inhibitors of gastric acid, were recorded, and the UBT was deferred for at least 1 month after any such medications were taken. Risk Factors for Reinfection There is a very high prevalence of H. pylori infection among family members of H. pylori infected children. 19,20 To determine whether the prevalence of infection within the family was a risk factor for reinfection, parents, siblings, and other household members were invited to undergo urea breath testing when the index child underwent the first reinfection assessment. Gastrointestinal symptoms and medication taken by siblings and parents were documented. Parents who had received therapy for the eradication of biopsy-proven H. pylori infection were identified. Details of family size, birth order of the index child, overcrowding, and socioeconomic status of the family were recorded. Overcrowding was present in a household if the crowding index (the number of persons in the house divided by the number of habitable rooms) was greater than 1 as defined by the Office of Population Census and Surveys. 21 Current family sleeping arrangements, including bed sharing, were documented. Children who currently slept with their parents more than 2 nights per week or shared a bed with a sibling were compared with those who slept in their own beds. The occupation of the highest income earner in the family was used to classify the socioeconomic status of the index child based on the Irish Social Class Scale. 22 The Irish Social Class Scale is an ordinal scale that includes 6 classes; class 1 is higher professional, and class 6 is unskilled manual workers. Families were categorized as upper socioeconomic group (classes 1 3) or lower socioeconomic group (classes 4 6, which include skilled, semiskilled, and unskilled manual workers). It was not possible to subdivide the lower socioeconomic group into individual classes because only 1 child became reinfected in socioeconomic class 5 and 1 in class 6. [ 13 C]Urea Breath Test The UBT was performed as described by Rowland et al. 18 Briefly, after a 2-hour fast, an oral dose of urea comprising 50 mg [ 13 C]urea or 75 mg [ 13 C]urea if body weight was 50 kg (99 atom%; Cambridge Isotopes, Cambridge, MA) was administered with 10 ml of a glucose polymer solution (Polycose; Abbott Laboratories, Ohio). Samples were collected before and 30 minutes after ingestion of the substrate. The ratio of 13 Cto 12 C in the baseline and 30-minute samples was determined by isotope ratio mass spectrometer (Europa Scientific ABACA, Brentford, England). Using this protocol, we have recently shown that an excess value of 3.5 (subtraction of baseline from 30-minute sample) is 100% sensitive and 97% specific for the diagnosis of infection in children. 18 Breath testing of adults and siblings older than 18 years was carried out as outlined by Braden et al. 23 None of the index children or siblings had an excess value close to the cutoff (i.e., between 3.0 and 4.5). One parent had an excess value of 3.7; this test was repeated and yielded an excess value of 1.0. Analysis of Samples All samples were analyzed at the Bureau of Stable Isotopes Analysis, Brentford, England. Statistical Analysis Infection rates were initially calculated per person-year of follow-up. Length of follow-up was determined from the date of confirmation of H. pylori eradication. The mean age of those who became reinfected was compared with that of those who remained clear of infection. Preliminary examination of the data indicated that there was a difference in age between the two groups; therefore, risk factors were stratified according to age. In a number of situations, stratified analysis became uninterpretable because of zero cells in cross-tabulations with age. The final group of variables considered in an analysis of risk factors included infection of both parents, infection of more than 2 older siblings, lower socioeconomic group, and bed sharing. Because of the variable length of follow-up (mean, months; range, 6 62 months), Cox regression analysis, as implemented in SAS (SAS Institute, Cary, NC) was used to estimate relative risks of reinfection. Forward stepwise methods were used to determine which variables were independently and significantly related to reinfection. Overall significance was set at the 5% level. For the purposes of risk factor analysis, parental infection

3 338 ROWLAND ET AL. GASTROENTEROLOGY Vol. 117, No. 2 was categorized as both parents currently infected or 1 or no parent currently infected. It was not possible to determine the H. pylori status of 2 mothers who were pregnant at the time of the study. In 1 of these families, the father was not infected; therefore, the family was categorized as having 1 or no parent infected. In the second family, the father was infected. Rather than exclude this family from multivariate analysis, we categorized the family as having both parents infected with H. pylori. This made no material difference to the presented results, which were similar to those obtained assuming the mother was not infected with H. pylori or excluding her altogether. Ethical Approval Ethical approval was obtained from the Ethics Committee of Our Lady s Hospital for Sick Children, Dublin, Ireland. Informed consent was obtained from the parents of children younger than 18 years of age. Informed consent was also obtained from parents and family members older than 18 years of age who were included in the study. Results Fifty-five of 60 eligible children participated in the study. Four children could not be contacted for follow-up, and 1 family refused to participate. There were 3 pairs of siblings in the study. All of these 6 children remained uninfected. To analyze risk factors, only 1 child from each family was included in the analysis. The oldest child was considered the index case, and results are therefore presented for 52 children. Details of the children in the study are given in Table 1. The mean duration of follow-up for the 52 index children to the last reinfection UBT was months (range, 6 62 months) (Table 1). Forty children (76.9%) were followed up for longer than 12 months, and 25 (48.1%) were followed up for more than 2 years. Table 1. Characteristics of Index Children and Their Families n 52 Mean age at treatment yr (range, yr) M:F ratio 1.6:1 Total patient-years of follow-up Mean duration of follow-up mo (range, 6 62 mo) Birth order 1st 21 (40.4%) 2nd 13 (25%) 3rd or greater 18 (34.6%) Mean siblings/family (range, 0 10) Mean infected siblings/family (range, 1 10) Mother infected 38 (77.5%) Father infected 33 (64.7%) Both parents infected 26 (50%) At least 1 infected sibling 34 (65.5%) Younger infected sibling 21 (40.4%) More than 2 older infected siblings 7 (13.5%) Lower socioeconomic group 25 (48.1%) Overcrowding index 1 15 (28.2%) Figure 1. Age at treatment of index children who remained noninfected or became reinfected with H. pylori. Six (11.5%) children became reinfected after patient-years of follow-up, and 46 of 52 children (88.5%) remained clear of infection. The overall reinfection rate was 5.8%/yr. The mean age of those who became reinfected was years (range, years). This compared with a mean age of years (range, years) for those who remained clear of infection (P ). Only 2 of 46 (4.3%) children who were older than 5 years of age became reinfected (Figure 1), compared with 4 of 6 children (66.6%) younger than 5 years. Four (33.3%) of 12 children younger than 10 years and 2 of 40 children older than 10 years became reinfected (Figure 1). The annual rate of reinfection was 2.0% per person-year of follow-up in children older than 5 years. Three of the 4 children younger than 5 years who became reinfected were treated with colloidal bismuth subcitrate, clarithromycin, and metronidazole for 1 week. 16 The fourth child was treated with colloidal bismuth subcitrate and amoxicillin. Of the 2 children older than 5 years who became reinfected, one was treated with colloidal bismuth subcitrate, tetracycline, and metronidazole for 4 weeks, and the other was treated with colloidal bismuth subcitrate, clarithromycin, and metronidazole for 1 week. Parents Fifty-one fathers and 49 mothers participated in the study. One father and 1 mother (different families) had died, and 2 mothers who were pregnant declined testing. Seventy-eight percent of mothers and 65% of fathers were currently infected (Table 1). Five (9.8%) fathers and 5 (10.2%) mothers who were not infected had received previous eradication therapy for biopsy-proven H. pylori infection. In 50% of families, both parents were currently infected; in 80.8% of families, at least 1 parent

4 August 1999 HELICOBACTER PYLORI REINFECTION IN CHILDREN 339 was infected. There were only 2 families in which no parent was currently infected. Siblings There were 148 siblings, and 139 (93.9%) of them underwent UBTs. Ninety-one siblings (65.5%) were infected with H. pylori (Table 1). Fifty percent of siblings (6 of 12) younger than 5 years were infected with H. pylori. The mean number of siblings was 3.8 ( 3.1) in families of children younger than 5 years old and 2.7 ( 1.82) in those of children older than 5 years (P 0.2). In 19 families (36.5%), all siblings were infected. In 15 (28.8%) families, only the index child was infected. Two index children had no siblings. Five of the 6 children (83%) younger than 5 years were currently sharing beds with parents or siblings, compared with 7 (15.2%) of those older than 5 years (P 0.001). Risk Factors for Reinfection Table 2 outlines how different risk factors considered in isolation relate to the risk of reinfection. Because analysis of the percentage of children reinfected does not take account of the variable follow-up, univariate relative risks as derived from Cox regression are presented. Using univariate analysis, infection of both parents and lower socioeconomic status were not risk factors for reinfection in children (Table 2). Age younger than 5 years, more than 2 older infected siblings, and bed sharing were risk factors for reinfection that reached statistical significance. Over the follow-up period, age was the strongest risk factor of those examined. All 5 factors were entered into a stepwise Cox regression analysis. Infection of both parents and lower socioeconomic status were again found not to be risk factors Table 2. Univariate Analysis of Risk Factors for Reinfection in Children Factor Factor present Reinfected/ n(%) Factor absent Reinfected/ n(%) Unadjusted RR (95% CI) Age 5 yr 4/6 (66.6) 2/46 (4.3) 22.0 ( ) Both parents Hp 5/27 (18.5) 1/25 (4.0) 4.5 ( ) 2 older siblings Hp 3/7 (42.8) 3/45 (6.6) 7.1 ( ) Lower socioeconomic group 5/25 (20.0) 1/27 (3.7) 6.0 ( ) Bed sharing a 4/13 (30.7) 2/39 (5.1) 7.2 ( ) NOTE. Cox regression analysis was used to estimate relative risk (RR) of reinfection associated with different risk factors considered in isolation while taking into account the variable length of follow-up for each participant in the study. 95% CI, 95% confidence interval; Hp, H. pylori positive. a Bed sharing with parent or sibling. Table 3. Regression Analysis Risk Factors Based on Age Adjusted for RR for age 5yr (95% CI) Significance of age effect (P) Not adjusted 22.0 ( ) Both parents Hp 18.5 ( ) older siblings Hp 16.1 ( ) Lower socioeconomic group 18.9 ( ) Bed sharing a 19.5 ( ) All 4 factors 15.0 ( ) NOTE. Stepwise regression analysis was used to examine the effect of confounders on the relationship of age and reinfection. The confounding effect of any single factor is small, and the effect of age on reinfection is independent of the other factors and remains strong and statistically significant. Each of the other factors was nonsignificant (P 0.4 in all cases). RR, relative risk; 95% CI, 95% confidence interval. a Bed sharing with parent or sibling. for reinfection. Furthermore, infection of more than 2 older infected siblings and bed sharing were not significant risk factors for reinfection in a regression analysis (P 0.4). Age was the only factor chosen that had a significant independent association with reinfection. To examine the possible confounding effect of the other factors on the relationship of age with reinfection, the relative risk associated with age less than 5 years was adjusted first for each factor separately and then for all factors simultaneously using Cox regression (Table 3). The results outlined in Table 3 show that the confounding effect of any single factor is small. The effect of age on reinfection is independent of the presence of the other factors and was statistically significant. When all 5 factors were considered simultaneously, age just failed to reach statistical significance (P 0.06), and none of the other factors was significant (P 0.4 in all cases). Discussion This is the first study to report on the risk of reinfection in children after successful treatment for H. pylori. Primary infection in adults is uncommon 13,24 ; therefore, it is not surprising that the reinfection rate in adults after treatment is approximately 1%/yr. 14,15,25,26 It is now accepted that H. pylori infection is almost always acquired in childhood. 11,13 Although the exact age of infection remains unknown, earlier studies suggest that infection occurs in preschool-aged children. Therefore, it is extremely important to determine the risk of reinfection in children because they are the group among whom infection is likely to be spread. The remarkable finding in this study is the very low rate of reinfection among children older than 5 years who were living in circumstances that would be considered to

5 340 ROWLAND ET AL. GASTROENTEROLOGY Vol. 117, No. 2 involve a high risk for primary infection. These factors include a high prevalence of infection among family members and poor socioeconomic conditions Eighty percent of families had 1 parent currently infected, and 65% of all siblings were infected. However, despite this very high prevalence of infection among family members, in regression analysis age was the only factor that was related to reinfection. Poor socioeconomic condition is the major risk factor for primary infection, 1,30 32 but it was not a risk factor for reinfection (Table 2). Overcrowding, which is also a risk factor for primary infection, 31,32 is highly correlated with poor socioeconomic conditions. Overcrowding could not be analyzed in this study because of zero cells in cross-tabulations with age. Our findings also suggest that children younger than 5 years may have a very high annual rate of reinfection. Although these findings in relation to children younger than 5 years of age are statistically significant, the number of children in this group is small, and further studies are needed to confirm this very high rate of reinfection in young children. The cutoff age of 5 years was chosen after initial analysis of the data clearly showed a statistically significant difference in age between those who became reinfected and those who remained clear of infection. Previous studies have shown that bed sharing in childhood is a risk factor for primary infection. 31,32 Using univariate analysis, current bed sharing was a significant risk factor for reinfection in this study (Table 2). However, our results suggest that bed sharing is a proxy for age, as shown by its nonsignificance in a full regression analysis (Table 3), rather than a specific risk factor for reinfection. In our study population, sharing a bed with a parent or a sibling is a common practice among young children; with 5 of the 6 children (83.3%) younger than 5 years currently sharing beds with family members, whereas only 15.5% of children older than 5 years were currently sharing beds with others. Recrudescence of infection is unlikely to explain our findings in children younger than 5 years. We included only children who had documented evidence of eradication 4 6 weeks after completion of treatment. Children younger than 5 years were treated according to the same treatment protocols as children older than 5 years. Because of their inclusion in a treatment study, 16 3ofthe 4 children younger than 5 years who became reinfected had 2 negative UBT results recorded, at 4 and again at 8 10 weeks after treatment, making recrudescence even more unlikely. One of the children was clear of infection at the initial assessment for reinfection 6 months after treatment but had become infected 10 months later. Length of follow-up in this study is variable. However, the children older than 5 years have a much longer follow-up (97.1 patient-years) than those younger than 5 years (6.75 patient-years), thereby reducing any potential bias. Furthermore, Cox regression analysis takes into account variable follow-up. Our finding may be important in relation to community screening and widespread treatment of H. pylori infected individuals. H. pylori has been classified as a group 1 carcinogen. 10 It remains unclear whether asymptomatic individuals infected with H. pylori should receive treatment to reduce the risk of gastric cancer. If widespread treatment of H. pylori infection is considered, it should be directed at individuals older than 5 years. Our results also suggest that treatment of family members is not necessary to prevent reinfection of patients treated for H. pylori associated duodenal ulcers. However, Brenner et al. 33 suggest that both a family history of peptic ulcer disease and infection with H. pylori are risk factors for peptic ulcer disease. Therefore, treatment of asymptomatic relatives may be considered. Our findings suggest that such treatment will lead to longterm clearance of infection. In conclusion, H. pylori reinfection is rare children older than 5 years, irrespective of poor socioeconomic conditions and a high prevalence of infection among family members. Reinfection in young children ( 5 years) may occur more frequently, and this will be the subject of further studies. The finding of a very low rate of reinfection in children older than 5 years may have significant implications for community screening and treatment of asymptomatic H. pylori infection. References 1. Peterson WL. Helicobacter pylori and peptic ulcer disease. N Engl J Med 1991;324: Dooley CP, Cohen H, Fitzgibbons PL, Bauer M, Appleman MD, Perez-Perez GI, Blaser MJ. Prevalence of Helicobacter pylori infection and histologic gastritis in asymptomatic persons. N Engl J Med 1989;321: Warren JR, Marshall BJ. Unidentified curved bacilli on gastric epithelium in active chronic gastritis. Lancet 1983;1: Blaser MJ. Gastric Campylobacter-like organisms, gastritis, and peptic ulceration. Gastroenterology 1987;93: Drumm B, Sherman P, Cutz E, Karmali M. Association of Campylobacter pylori on the gastric mucosa with antral gastritis in children. N Engl J Med 1987;316: Walsh JH, Peterson WL. The treatment of Helicobacter pylori infection in the management of peptic ulcer disease. N Engl J Med 1995;333: Hentschel E, Brandstatter G, Dragosics B, Hirschl AM, Nemec H, Taufer M, Wurzer H. Effect of ranitidine and amoxicillin plus metronidazole on the eradication of Helicobacter pylori and the recurrence of duodenal ulcer. N Engl J Med 1993;328: Coughlan JG, Gilligan D, Humphries H, McKenna D, Dooley C, Sweeney E, Keane C, O Morain C. Campylobacter pylori and recurrence of duodenal ulcers a 12 month follow up study. Lancet 1987;2:

6 August 1999 HELICOBACTER PYLORI REINFECTION IN CHILDREN Graham DY, Lew GM, Klein PD, Evans DG, Evans DJ Jr, Saeed ZA, Malaty HM. Effect of treatment of Helicobacter pylori infection on the long-term recurrence of gastric and duodenal ulcer. A randomized, controlled study. Ann Intern Med 1992;116: Monographs on the evaluation of carcinogenic risks to humans. Schistosomes, liver flukes and Helicobacter pylori. Lyon, France: International Agency for Research on Cancer, 1994: Banatvala N, Mayo K, Megraud F, Jennings R, Deeks JJ, Feldman RA. The cohort effect and Helicobacter pylori. J Infect Dis 1993;168: Blaser MJ, Chyou PH, Nomura A. Age at establishment of Helicobacter pylori infection and gastric carcinoma, gastric ulcer, and duodenal ulcer risk. Cancer Res 1995;55: Cullen DJ, Collins BJ, Christiansen KJ, Epis J, Warren JR, Surveyor I, Cullen KJ. When is Helicobacter pylori infection acquired? Gut 1993;34: Borody T, Andrews P, Mancuso N, Jankiewicz E, Brandl S. Helicobacter pylori reinfection 4 years post-eradication (letter). Lancet 1992;339: Forbes GM, Glaser ME, Cullen DJ, Warren JR, Christiansen KJ, Marshall BJ, Collins BJ. Duodenal ulcer treated with Helicobacter pylori eradication: seven-year follow-up. Lancet 1994;343: Walsh D, Goggin N, Rowland M, Durnin M, Moriarty S, Drumm B. One week treatment for Helicobacter pylori infection. Arch Dis Child 1997;76: Gormally SM, Prakash N, Durnin MT, Daly LE, Clyne M, Kierce BM, Drumm B. Association of symptoms with Helicobacter pylori infection in children. J Pediatr 1995;126: Rowland M, Lambert I, Gormally S, Daly LE, Thomas JE, Hetherington CD, Drumm B. Carbon 13 labeled urea breath test for the diagnosis of Helicobacter pylori in children. J Pediatr 1997;131: Drumm B, Perez-Perez GI, Blaser MJ, Sherman PM. Intrafamilial clustering of Helicobacter pylor infection. N Engl J Med 1990;322: Malaty HM, Graham DY, Klein PD, Evans DG, Adam E, Evans DJ. Transmission of Helicobacter pylori infection. Studies in families of healthy individuals. Scand J Gastroenterol 1991;26: Office of Population Censuses and Surveys. Britain s households. Census guide. Volume 4. 4th ed. London: Her Majesty s Stationary Office, 1986: O Hare A, Whelan CT, Commins P. The development of an Irish census-based social class scale. Econ Soc R 1991;22: Braden B, Duan L P, Caspary W P, Lembcke B. More convenient 13 C-urea breath test modifications still meet the criteria for valid diagnosis of Helicobacter pylori infection. Z Gastroenterol 1994; 32: Kuipers EJ, Pena AS, van Kamp G, Uyterlinde AM, Pals G, Pels NF, Meuwissen SG. Seroconversion for Helicobacter pylori. Lancet 1993;342: van der Hulst RW, Rauws EA, Koycu B, Keller JJ, ten Kate FJ, Dankert J, Tytgat GN, van der Ende A. Helicobacter pylori reinfection is virtually absent after sucessful eradication. J Infect Dis 1997;1: Mitchell HM, Hu P, Chi Y, Chen MH, Li YY, Hazell SL. A low rate of reinfection following effective therapy against Helicobacter pylori in a developing nation (China). Gastroenterology 1998;114: Perez-Perez GI, Taylor DN, Bodhidatta L, Wongsrichanalai J, Baze WB, Echeverria PD, Blaser MJ. Seroprevalence of Helicobacter pylori infections in Thailand. J Infect Dis 1990;161: Lindkvist P, Asrat D, Nilsson I, Tsega E, Olsson GL, Wretlind B, Giesecke J. Age at acquisition of Helicobacter pylori infection: comparison of a high and low prevalence country. Scand J Infect Dis 1996;28: Mitchell HM, Li YY, Hu PJ, Liu Q, Chen M, Du GG, Wang ZJ, Lee A, Hazell SL. Epidemiology of Helicobacter pylori in Southern China: identification of early childhood as the critical period for acquisition. J Infect Dis 1992;166: Fiedorek SC, Malaty HM, Evans DL, Pumphrey CL, Casteel HB, Evans DJJ, Graham DY. Factors influencing the epidemiology of Helicobacter pylor infection in children. Pediatrics 1991;88: McCallion WA, Murray LJ, Bailie AG, Dalzell AM, O Reilly DP, Bamford KB. Helicobacter pylor infection in children: relation with current household living conditions. Gut 1996;39: Webb PM, Knight T, Greaves S, Wilson A, Newell DG, Elder J, Forman D. Relation between infection with Helicobacter pylori and living conditions in childhood: evidence for person to person transmission in early life. BMJ 1994;308: Brenner H, Rothenbacher D, Bode G, Adler G. The individual and joint contributions of Helicobacter pylori infection and family history to the risk for peptic ulcer disease. J Infect Dis 1998;117: Received December 31, Accepted April 20, Address requests for reprints to: Brendan Drumm, M.D., Children s Research Centre, Our Lady s Hospital for Sick Children, Crumlin, Dublin 12, Ireland. paeds@crumlin.ucd.ie; fax: (353) Supported by a Wellcome Trust Project grant and by BIOMED CT The authors thank Nurse Marie Durnin and Sister Anna Lloyd for their help in organizing this project; Philip Johnson of Bureau of Stable Isotopes Analysis, United Kingdom, for his support in carrying out urea breath tests; and Dr. Billy Bourke for his critical review of the manuscript.