Effect of sibling number in the household and birth order on prevalence of Helicobacter pylori: a cross-sectional study

Transcription

1 Published by Oxford University Press on behalf of the International Epidemiological Association ß The Author 2007; all rights reserved. Advance Access publication 28 September 2007 International Journal of Epidemiology 2007;36: doi: /ije/dym201 Effect of sibling number in the household and birth order on prevalence of Helicobacter pylori: a cross-sectional study Alexander C Ford, 1 * David Forman, 2 Alastair G Bailey, 1 Karen J Goodman, 3 Anthony T R Axon 1 and Paul Moayyedi 4 Accepted 29 August 2007 Background Methods Results Infection with Helicobacter pylori (H. pylori) is acquired mainly in childhood, with studies demonstrating this is related to living conditions. Effects of sibling number and birth order on prevalence of infection have not been extensively studied. The authors performed a cross-sectional survey of adults, aged between 50 and 59 years, previously involved in a community-screening programme for H. pylori in Leeds and Bradford, UK. Prevalence of H. pylori was assessed at baseline with urea breath test. All individuals who were alive, and could be traced, were contacted by postal questionnaire in 2003 obtaining information on number of siblings and birth order. Data concerning childhood socioeconomic conditions were stored on file from the original study (47%) of 8407 original participants provided data. Prevalence of infection increased according to sibling number (20% in those with none vs 63% with eight or more). Controlling for childhood socioeconomic conditions and birth order using multivariate logistic regression, infection odds were substantially increased with three siblings compared with none [odds ratio (OR) 1.51; 95% confidence interval (CI) ], and a gradient of effect continued up to eight or more siblings (OR 5.70; 95% CI ). Odds of infection also increased substantially with birth order, but the positive gradient disappeared on adjustment for sibling number and childhood socioeconomic conditions. Conclusions: In this cross section of UK adults, aged years, sibling number in the household, but not birth order, was independently associated with prevalence of H. pylori infection. Keywords Birth order, Helicobacter pylori, prevalence, siblings Introduction Helicobacter pylori (H. pylori) is a spiral-shaped gram-negative bacillus that infects the human stomach. 1 The organism is the 1 Centre for Digestive Diseases, Leeds General Infirmary, Great George Street, Leeds, UK. 2 Centre for Epidemiology and Biostatistics, Medical School, Leeds University, Leeds, UK. 3 Department of Medicine and School of Public Health, University of Alberta, Edmonton, Alberta, Canada. 4 Gastroenterology Division, McMaster University, Health Sciences Centre, Hamilton, Ontario, Canada. * Corresponding author. Room 7.23, Department of Academic Medicine, Clinical Sciences Building, St. James s University Hospital, Leeds, LS9 7TF, England, UK. alexf12399@yahoo.com main cause of peptic ulcer disease, and is strongly associated with distal gastric adenocarcinoma. 2,3 The prevalence of infection varies according to the geographic region under study, and the demographics of the sampled population within that region. 4 6 In general, the prevalence is lower in developed countries, and this may be as a result of improvements in living conditions in the last century. Acquisition of H. pylori infection is thought to occur predominantly in childhood. 7,8 Currently, the majority of evidence points to transmission from human-to-human, with studies suggesting that intra-familial transmission is a major source of infection However, the mode of transmission remains unknown. Several have been proposed including faeco oral, oro oral and gastro oral, and the 1327

2 1328 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY bacterium has been successfully cultured from stool, saliva and vomitus Factors that may influence the transmission of H. pylori have been studied previously. Infection with the bacterium appears to be primarily related to living conditions in childhood. The degree of household overcrowding, sharing a bed or bedroom with another sibling as a child, father s social class and lack of hot water have all been demonstrated to have a positive association with infection in some populations We, and others, have previously demonstrated a positive relationship between prevalence of H. pylori and number of other siblings in the household during childhood There are, however, few studies that have examined the relationship between order of birth among siblings and infection with the bacterium. One large study demonstrated that higher birth order was associated with increased prevalence of infection in children, 24 suggesting that older children pass on the infection to their younger siblings, but this study was conducted in a Colombian population, and therefore the findings may not be generalizable to settings with higher levels of socioeconomic development. We have examined the relationship between sibling number, birth order and prevalence of H. pylori in a cross-sectional study of a large number of adult individuals nested within a 10-year longitudinal survey of the natural history of dyspepsia. Methods Study design and sample The study was a cross-sectional survey of individuals, now aged between 50 and 59 years, who were randomly selected from the general population and recruited into a randomized controlled trial of H. pylori screening and treatment, conducted in Leeds and Bradford, both cities in West Yorkshire, Northern England, with a combined population of approximately The original study commenced in 1994, with screening for H. pylori via 13-carbon urea breath test. It became apparent, following analysis of baseline data from the original study, that number of siblings was an important risk factor for infection with H. pylori. 23 We therefore assessed the influence of birth order at 10-year follow up, in an attempt to determine whether we could observe evidence of transmission from older to younger siblings during childhood as a predictor of prevalent infection in middle-aged adults. We attempted to contact all 8407 original participants at 10 years, regardless of H. pylori status, as part of a longitudinal survey examining the natural history of dyspepsia. This was achieved by visiting the 36 primary care centres that had originally recruited the individuals, to obtain a current address for correspondence from computer databases. Those who were alive and traceable to a current address were sent a validated dyspepsia questionnaire by post. This also collected information regarding the number of other siblings in the household during childhood, as well as their order of birth. The latter two questions were left open to interpretation by the respondent, in that we did not ask how many siblings lived within the same household at a particular age, nor did we specifically instruct the subjects to include deceased, half, or step siblings. Non-responders were sent a second postal questionnaire. The relevant local research ethics committees in Leeds and Bradford approved the study in March 2003, and the initial questionnaires were sent out in August Baseline demographic data and socioeconomic conditions in childhood at 8 years of age including: gender; presence of inside toilet and bathroom; shared bedroom; shared bed with another sibling; type of housing (flat, back-to-back house, terraced house, or detached/semi-detached house); household crowding (number of occupants/room); father s social class (as defined by occupation) and H. pylori status were available for all individuals from the original study. siblings was also collected at baseline, but these data were not used for these analyses. However, there was a high degree of agreement (97%) between the numbers of siblings reported by individuals at both time points. Statistical analysis We compared demographic data and data concerning socioeconomic conditions in childhood between responders and those who could not be traced to a current address or did not respond using the 2 -test. The associations between number of siblings in the household, birth order and prevalence of H. pylori infection were estimated as odds ratios (ORs) with 95% confidence intervals (CIs). The influence of gender, presence of inside toilet and bathroom, shared bedroom, shared bed with another sibling, type of housing, household crowding and father s social class on the relationship between number of siblings, birth order and H. pylori infection were investigated using a logistic regression model. All P-values were two-tailed. All statistical analyses were performed using StatsDirect statistical software, version (StatsDirect Ltd., Sale, Cheshire, UK) and SPSS for Windows, version 14.0 (SPSS Inc., Chicago, IL, USA). Results Of the 8407 participants originally enrolled, 1991 (24%) could not be traced to a current address. Questionnaires were sent to the remaining 6416, and 4003 subjects (48%) responded (Figure 1). Of these, 75 did not respond to the question concerning number of siblings in the household during childhood, leaving 3928 (47%) with analysable data. In addition, 23 individuals did not respond to the question concerning order of birth; therefore 3905 (46%) individuals provided data for this analysis. The prevalence of H. pylori infection in those who responded to the questionnaire was 27% (1086 of 4003 individuals), and 44% of responders were male (1756 individuals). Differences in demographic data and in socioeconomic conditions in childhood between responders and non-responders are shown in Table 1. We were more likely to successfully contact females, those whose father was of higher social class, and those with no siblings. Those who shared a bedroom with another sibling during childhood were less likely to be successfully followed up. The prevalence of H. pylori was nearly equal in the two groups. Of the 3928 subjects that completed the question concerning sibling number, 605 (15%) were only children. Those who were H. pylori-positive had a mean of 2.71 siblings, compared with

3 EFFECT OF SIBLING NUMBER IN HOUSEHOLD ON H. PYLORI PREVALENCE originally enrolled 6416 traced and sent questionnaire 4003 returned questionnaire 3928 provided fully analysable data Figure 1 Flow of study participants 1991 no contact address 2413 did not respond 75 questionnaires incomplete Table 1 Demographics and socioeconomic conditions in childhood (at 8 years of age) of individuals who were successfully contacted compared to those who were not contacted or did not respond Responded (n ¼ 4003) Not contacted or no response (n ¼ 4404) P-value Gender Male 1756 (44%) 2180 (49.5%) < Female 2247 (56%) 2224 (50.5%) Father s social class I and II 722 (18%) 689 (16%) III 2558 (64%) 2800 (64%) IV and V 454 (11%) 573 (13%) H. pylori positive 1086 (27%) 1243 (28%) 0.32 siblings (15%) 512 (12%) < (31%) 1285 (29%) (23%) 1012 (23%) (12%) 550 (12.5%) (7%) 359 (8%) (4%) 247 (6%) 6 97 (2%) 120 (3%) 7 49 (1%) 103 (2%) (2%) 154 (3.5%) Inside WC 3023 (75.5%) 3306 (75%) 0.68 Bathroom 751 (19%) 834 (19%) 0.81 Shared bedroom as a child 2219 (55%) 2575 (58.5%) Shared bed as a child 952 (24%) 1080 (24.5%) 0.41 occupants per room (4%) 161 (4%) (63%) 2586 (59%) (31%) 1572 (36%) Mean age in years (SD) 55.4 (2.8) 55.2 (2.9) in H. pylori-negative individuals (P < ). The prevalence of H. pylori infection increased steadily with increasing number of siblings in the household (Table 2), except for a small decrease at six, which is likely a random fluctuation considering the prevalence CIs. In those with no siblings, the prevalence of infection was 20%, compared with 63% in those with eight or more siblings. There was a substantial increase in the odds of infection with three or more siblings compared with none, (OR ¼ 1.79; 95% CI: ), and a gradient of effect continued up to eight or more siblings (OR ¼ 6.69; 95% CI: ). This trend remained strong, though somewhat reduced, following logistic regression to control for gender and socioeconomic conditions in childhood. When gender, socioeconomic conditions in childhood and birth order were controlled for, the trend retained a similar gradient. Modelling the number of siblings in continuous rather than categorical form, the unadjusted OR per increase of each additional sibling was 1.26 (95% CI: , P < ), and this increase remained similar in magnitude following logistic regression controlling for gender, socioeconomic conditions in childhood and birth order (adjusted OR 1.23; 95% CI: , P < ). When examining the effect of birth order on prevalence of H. pylori, the odds of infection became substantially higher when the individual had two older siblings compared with being first born (OR ¼ 1.46; 95% CI: ), and this increased steadily with increasing number of older siblings, up to six or more (OR ¼ 4.16; 95% CI: ) (Table 3). This relationship weakened following logistic regression to adjust for gender and socioeconomic conditions in childhood, though the odds of infection remained substantially elevated for five older siblings or more, suggesting that some of the observed effect was due to confounding by other socioeconomic conditions in childhood. When number of siblings was also controlled for using logistic regression, the positive association completely disappeared across all birth order strata. When the relationship between prevalence of H. pylori infection and total number of siblings in the household was examined for first-born individuals only, there remained a positive gradient of association between the two, with

4 1330 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY Table 2 Total number of siblings and prevalence of H. pylori infection siblings in the household individuals (n ¼ 3928) Number H. pyloripositive (%) (n ¼ 1058) Odds ratio ratio a ratio (95% confidence (20) (21) 1.08 ( ) 1.02 ( ) 1.02 ( ) (24) 1.27 ( ) 1.17 ( ) 1.19 ( ) (31) 1.79 ( ) 1.46 ( ) 1.51 ( ) (36) 2.21 ( ) 1.66 ( ) 1.85 ( ) (45.5) 3.31 ( ) 2.50 ( ) 2.72 ( ) (41) 2.78 ( ) 2.00 ( ) 2.33 ( ) (55) 4.86 ( ) 3.07 ( ) 3.47 ( ) (63) 6.69 ( ) 4.69 ( ) 5.70 ( ) a Adjusted for gender, inside toilet, bathroom, shared bedroom, shared bed with other siblings, type of housing, household crowding as a child, and father s social class ( adjusted for all of the above plus birth order). Table 3 Birth order and prevalence of H. pylori infection Birth order of individual individuals (n ¼ 3905) Number H. pyloripositive (%) (n ¼ 1054) Odds ratio ratio a ratio 1st (24) nd (26) 1.12 ( ) 1.04 ( ) 1.01 ( ) 3rd (31) 1.46 ( ) 1.15 ( ) 0.96 ( ) 4th (34) 1.64 ( ) 1.25 ( ) 0.90 ( ) 5th (35) 1.75 ( ) 1.13 ( ) 0.70 ( ) 6th (48) 2.97 ( ) 1.67 ( ) 0.82 ( ) 7th to 16th (56) 4.16 ( ) 2.42 ( ) 0.82 ( ) a Adjusted for gender, inside toilet, bathroom, shared bedroom, shared bed with other siblings, type of housing, household crowding as a child, and father s social class ( adjusted for all of the above plus sibling number). substantially higher odds of infection in those with three or more siblings compared with none (OR of infection with three siblings ¼ 1.77; 95% CI: ) (Table 4). This relationship persisted in a logistic regression model adjusting for gender and childhood socioeconomic conditions. Given that individuals of higher birth order cannot have a small number of siblings, there is likely to be imperfect control for confounding in the previous analyses. For this reason, the relationship between sibling number, birth order and prevalence of H. pylori infection was examined further by performing analyses of the effect of birth order as a continuous variable within strata of sibling numbers (from one to seven or more), and the effect of sibling number as a continuous variable within birth order strata (from first to seventh or later), (Table 5) using multivariate logistic regression to control for gender and childhood socioeconomic conditions. Birth order had near-null associations with prevalence of H. pylori infection within sibship-size strata, whilst sibling number continued to display a positive trend within birth order strata. For birth orders between 1st and 5th, increases in infection odds with each additional sibling ranged from 11 to 26%, and the trend became stronger among those with five older siblings, though the precision weakened at higher birth orders due to smaller numbers. Discussion This large cross-sectional survey has demonstrated that the prevalence of H. pylori infection in UK adults in the north of England, aged between 50 and 59 years, is closely related to the number of other siblings in the household during childhood, even after adjustment for socioeconomic conditions in childhood and birth order. The order of the individual s birth within the family also appeared strongly related to the prevalence of infection with the bacterium, but this association became near null when other possible confounding socioeconomic conditions in childhood were controlled for within strata of sibship size in a logistic regression model, suggesting that the apparent increase in odds of H. pylori infection in those who were not first born was predominantly due to these individuals having more siblings. The fact that the relationship between number of siblings and prevalence of H. pylori remained within strata of birth order, whilst birth order had little effect within any strata of sibship size, provides further evidence that the latter is a less important influence on the prevalence of infection in adult life. There are several studies that have examined these issues previously, but few within a large data set such as this, with individuals who were originally randomly selected from the

5 EFFECT OF SIBLING NUMBER IN HOUSEHOLD ON H. PYLORI PREVALENCE 1331 Table 4 siblings in the household and prevalence of H. pylori infection in first-born individuals other siblings in the household first-born individuals (n ¼ 1747) Number H. pylori-positive (%) (n ¼ 414) Odds ratio ratio a (20) (22) 1.12 ( ) 0.98 ( ) (23) 1.21 ( ) 1.07 ( ) (31) 1.77 ( ) 1.37 ( ) (38) 2.46 ( ) 2.02 ( ) (42) 2.92 ( ) 2.34 ( ) (50) 3.96 ( ) 2.24 ( ) (58) 5.54 ( ) 3.31 ( ) a Adjusted for gender, inside toilet, bathroom, shared bedroom, shared bed with other siblings, type of housing, household crowding as a child, and father s social class. Table 5 Effect of birth order within strata of sibship size and effect of number of siblings within birth order strata on prevalence of H. pylori siblings ratio a per unit increase in birth order (95% confidence Birth order ratio a per unit increase in siblings ( ) 1st 1.15 ( ) ( ) 2nd 1.26 ( ) ( ) 3rd 1.20 ( ) ( ) 4th 1.11 ( ) ( ) 5th 1.12 ( ) ( ) 6th 1.93 ( ) ( ) 7th to 16th 1.23 ( ) a Adjusted for gender, inside toilet, bathroom, shared bedroom, shared bed with other siblings, type of housing, household crowding as a child, and father s social class. general population, and which is therefore not prone to the inherent biases of studies carried out in primary or secondary care populations. In a study of 65- to 75-year-old individuals, conducted in Hertfordshire, England, prevalence of H. pylori infection was determined by serological analysis, and factors in childhood that may influence seropositivity were examined. 20 As in our study, the authors reported that prevalence of H. pylori rose with increasing number of siblings, and with increasing order of birth, though the latter association disappeared following multivariate logistic regression. In a German study of adult patients attending their primary care physician who were screened for H. pylori by urea breath test, without regard to the reason for consultation, a positive association between the number of siblings and prevalence of infection was demonstrated, though this was not a populationbased study like ours. 19 The same authors have also examined this association in children attending a pre-school medical examination. 21 Again, H. pylori status was assessed by urea breath test. This study reported that number of siblings was only a weak predictor of positivity, and an association with birth order was not demonstrated, though the prevalence of infection in this study was only 13%, which is lower than in our study, and few children had more than two siblings. In a rural population in the Colombian Andes with one of the highest incidences of gastric carcinoma in the world, a census sample of 684 local children aged between 2 and 9 years of age were screened for H. pylori with urea breath test. 24 The authors reported an independent association between both sibling number and birth order and infection with the bacterium. In contrast to the findings of our study, birth order remained associated with infection following logistic regression to control for number of other children in the family, though the association was not as strong. This study also examined age gap to next oldest sibling, and demonstrated that children born within 4 years of an older sibling were more likely to be H. pylori positive. Possible explanations for the differing result of this report and the current one are that the Colombian study was examining this association in children rather than in adults. It may be that in Colombia at the age when infection is primarily acquired, the observed relationship with birth order is directly reflective of the route of transmission from older siblings to younger, or perhaps transient infection with H. pylori occurs in childhood, and predictors of onset of infection may differ from predictors of persistence into adult life. Alternatively, the contrast in results may be due to socioeconomic differences between the UK and Colombia. In communities where age gaps between siblings are wider, contact with other children outside the home may be more important in determining acquisition of infection than contact with siblings. 26 In addition, household hygiene standards may be more dependent on family size in the UK than Colombia, and this may explain why sibship size dominates any effect of birth order in the UK. A Japanese population survey conducted amongst schoolchildren screened for presence of H. pylori with salivary antigen, and failed to demonstrate an effect of either birth order or sibship size on prevalence of infection, but the number of infected children was small, and the low sensitivity of the salivary antigen test (84%) may have produced nondifferential misclassification that masked the association. 27 However, another Japanese population-based study of over 5000 adult individuals suggested that later birth order was associated with infection. 28 More recently, investigators from Greenland demonstrated that both total number of siblings and number of older siblings were associated with prevalence of H. pylori. 22

6 1332 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY There are some limitations of this study, which are imposed by its design. The high attrition rate, with only 47% of those originally involved being successfully traced, responding to the questionnaire and providing fully analysable data is almost inevitable when following up a large number of individuals randomly selected from the general population after such a long period of time. There were some differences between those successfully contacted and those who were lost to follow up, but it is unlikely that this would be responsible for the different trends estimated for sibship size and birth order. In addition, the restricted age range of these individuals, between 50 and 59 years at the time this study was conducted, means that the results may not be generalizable to all age cohorts. Whilst there is some vulnerability to recall bias in studies examining retrospective issues such as these, it is unlikely that an individual would be unable to correctly report how many siblings they had. There are some complex issues that relate to how a sibling is defined, including whether to include deceased siblings and half or step siblings but, in general, deviations from the typical nuclear family amongst this large sample of individuals are probably not frequent enough to have caused any significant amount of misclassification. This study also has several strengths. Despite the high losses to follow up, it is a very large study, with almost 4000 individuals providing data. It is therefore one of the largest conducted that has examined the relationship between sibling number, birth order and H. pylori infection, whilst simultaneously controlling for other potential confounding socioeconomic factors in childhood. Finally, the use of the urea breath test to determine H. pylori status is another strong point, as this technique has an estimated sensitivity and specificity of 95% for the diagnosis of infection with the bacterium. 29 In conclusion, we have demonstrated in a large crosssectional survey that the prevalence of infection with H. pylori, between ages 50 and 59 years, is strongly related to the number of siblings in the household during childhood. This observation remained stable following multivariate logistic regression to control for other potential confounding factors. Birth order appeared to be a less important determinant, when the number of other siblings in the household was controlled for. These findings add further support to the hypothesis that transmission of infection is from person-to-person, and occurs predominantly in childhood, but it would appear that, in populations such as the one studied, birth order does not influence or predict prevalence of H. pylori in later life. Conflict of interest: David Forman has received speakers and consulting fees from AstraZeneca, Wyeth and Takeda. Paul Moayyedi has received speaker s fees and research funds from AstraZeneca, Wyeth Laboratories and Abbott Laboratories. The remaining authors declared none. Sources of funding: None declared. KEY MESSAGES H. pylori is acquired in childhood and this is related to living conditions. Previous studies conducted in the developed world have demonstrated a significant relationship between prevalence of H. pylori and sibling number in the household during childhood. Studies from the developing world have suggested that later birth order pre-disposes to infection with the bacterium, but this issue has not been extensively studied in the developed world. This large UK population-based study confirmed that prevalence of H. pylori infection in adult life was independently associated with sibling number in the household during childhood, but not birth order in the household. References 1 Warren JR, Marshall BJ. Unidentified curved bacilli on gastric epithelium in active chronic gastritis. Lancet 1983;321: International Agency for Research on Cancer. Schistosomes, liver flukes and Helicobacter pylori. IARC Monogr Eval Carcinog Risks Hum 1994;61: NIH Consensus Development Panel on Helicobacter pylori in Peptic Ulcer Disease. Helicobacter pylori in peptic ulcer disease. JAMA 1994;272: Mitchell HM, Li YY, Hu PJ et al. Epidemiology of Helicobacter pylori in Southern China: identification of early childhood as the critical period for acquisition. J Infect Dis 1992;166: Perez-Perez GI, Taylor DN, Bodhidatta L et al. Seroprevalence of Helicobacter pylori infections in Thailand. J Infect Dis 1990;161: Graham DY, Malaty HM, Evans DG, Evans DJ, Klein PD, Adam E. Epidemiology of Helicobacter pylori in an asymptomatic population in the United States. Gastroenterology 1991;100: Banatvala N, Mayo K, Megraud F, Jennings R, Deeks JJ, Feldman RA. The cohort effect and Helicobacter pylori. J Infect Dis 1993;168: Nabwera HM, Nguyen-Van-Tam JS, Logan RFA, Logan RPH. Prevalence of Helicobacter pylori infection in Kenyan schoolchildren aged 3-15 years and risk factors for infection. Eur J Gastroenterol Hepatol 2000;12: Drumm B, Perez-Perez GI, Blaser MJ, Sherman PM. Intrafamilial clustering of Helicobacter pylori infection. N Engl J Med 1990;322: Mitchell HM, Bohane TD, Berkowicz J, Hazell SL, Lee A. Antibody to Campylobacter pylori in families of index children with gastrointestinal illness due to C. pylori. Lancet 1987;330: Parente F, Maconi G, Sangaletti O et al. Prevalence of Helicobacter pylori infection and related gastroduodenal lesions in spouses of Helicobacter pylori positive patients with duodenal ulcer. Gut 1996;39:

7 EFFECT OF SIBLING NUMBER IN HOUSEHOLD ON H. PYLORI PREVALENCE 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: Thomas JE, Gibson GR, Darboe MK, Weaver LT, Dale A. Isolation of Helicobacter pylori from human faeces. Lancet 1992;340: Leung WK, Siu KLK, Kwok CKL, Chan SY, Sung R, Sung JJY. Isolation of Helicobacter pylori from vomitus in children and its implication in gastro-oral transmission. Am J Gastroenterol 1999;94: Parsonnet J, Shmuely H, Haggerty T. Fecal and oral shedding of Helicobacter pylori from healthy infected adults. JAMA 1999;282: Galpin OP, Whitaker CJ, Dubiel AJ. Helicobacter pylori infection and overcrowding in childhood. Lancet 1992;339: Mendall MA, Goggin PM, Molineaux N et al. Childhood living conditions and Helicobacter pylori seropositivity in adult life. Lancet 1992;339: Webb PM, Knight T, Greaves S et al. Relation between infection with Helicobacter pylori and living conditions in childhood: evidence for person to person transmission in early life. Br Med J 1994;308: Rothenbacher D, Bode G, Winz T, Berg G, Adler G, Brenner H. Helicobacter pylori in out-patients of a general practitioner: prevalence and determinants of current infection. Epidemiol Infect 1997;119: Fall CDH, Goggin PM, Hawtin P, Fine S, Duggleby S. Growth in infancy, infant feeding, childhood living conditions, and Helicobacter pylori infection at age 70. Arch Dis Child 1997;77: Rothenbacher D, Bode G, Berg G et al. Prevalence and determinants of Helicobacter pylori infection in preschool children: a populationbased study from Germany. Int J Epidemiol 1998;27: Koch A, Krause TG, Krogfelt K, Olsen OR, Fischer TK, Melbye M. Seroprevalence and risk factors for Helicobacter pylori infection in Greenlanders. Helicobacter 2005;10: Moayyedi P, Axon ATR, Feltbower R et al. Relation of adult lifestyle and socioeconomic factors to the prevalence of Helicobacter pylori infection. Int J Epidemiol 2002;31: Goodman KJ, Correa P. Transmission of Helicobacter pylori among siblings. Lancet 2000;355: Moayyedi P, Forman D, Braunholtz D et al. The proportion of upper gastrointestinal symptoms in the community associated with Helicobacter pylori, lifestyle factors, and nonsteroidal anti-inflammatory drugs. Am J Gastroenterol 2000;95: Patel P, Mendall MA, Khulusi S, Northfield TC, Strachan DP. Helicobacter pylori infection in childhood: risk factors and effect on growth. Br Med J 1994;309: Kurosawa M, Kikuchi S, Inaba Y, Ishibashi T, Kobayashi F. Helicobacter pylori infection, gastritis and gastric cancer. J Gastroenterol Hepatol 2000;15: Ueda M, Kikuchi S, Kasugai T, Shunichi T, Miyake C. Helicobacter pylori risk associated with childhood home environment. Cancer Sci 2003;94: Logan RPH, Dill S, Bauer FE et al. The European 13 C-urea breath test for the detection of Helicobacter pylori. Eur J Gastroenterol Hepatol 1991;3: