This is the first survey describing the prevalence of Helicobacter pylori colonization among apparently healthy children in an urban area in Uganda. The study revealed a high overall prevalence rate of Helicobacter pylori, 44.3% and early colonization was common. Boys were significantly more often colonized than girls and those living in semi-permanent houses more often than those living in permanent houses.
The strengths of our study are a) a test method with high sensitivity, specificity and accuracy in comparison to other methods [21, 22] b) the study population is representative for this community; c) few potential participants declined to participate (6.6%) and few potential participants were excluded, (3.1%). Only 5 children were excluded from the final analysis due to HIV infection and any selection bias is minimal. We made efforts to ensure that our study population was healthy, and therefore it can be used as reference population for future studies.
However, we have not controlled for Helicobacter pylori colonization in the caretakers or siblings. This could potentially bias our prevalence estimate upwards. High resolution molecular methods used recently in South-Africa [29, 30] to resolve transmission of Helicobacter pylori are highly invasive which limits their feasibility in low-resource settings.
Our findings are comparable to findings from other sub-Saharan countries [5, 31]. Very few studies carried out in sub-Saharan Africa have used the accurate, non invasive method as we did . A study carried out in apparently healthy Kenyan children, using serological tests found a prevalence of 45.6% . A study from Cameroon  using serological tests had results similar to ours with a prevalence of 37.5% in children younger than 3 years. We found an increasing prevalence with age. These findings are comparable with findings from Cameroon, Nigeria, Gambia and Egypt [7, 9, 13, 33]. We detected a lower prevalence in children aged 9- < 12 years compared to 6- < 9 years. This finding is not comparable with, for instance, the prevalence found in the Ugandan study from Newton et al  in adults, where the prevalence of Helicobacter pylori in adults suffering from different kinds of cancer, except gastric cancer, was 87%. Our study population was distinctly different as all our participants were apparently healthy and children. A study from Iran also found a lower prevalence in children older than 14 years . In the era before the stool antigen test became available, several studies also found a decrease in the prevalence with age [36, 37], suggesting spontaneous eradication, better attention to health issues in older children, or use of antibiotics for other common diseases [36, 39]. Another explanation of this finding could be an increasing antibody production with increasing age that may lead to the decline of the prevalence rate in older children [36, 37]. Differences in types of Helicobacter pylori in adults compared to children, and differences in special gastric receptors have been suggested as other explanations for this decrease in prevalence . Auto-curability among black children, age 7-21 years in USA was found to be 0.3% per year and 5.5% per year among white children in the same cohort . Among Peruvian children, a spontaneous eradication of 7% per month was found (6-30 months old) . In the youngest children age 0 < 6 months we found a high prevalence. A high prevalence has been found in neonates , decreasing in older babies and toddlers, suggesting an auto-curability and that acquisition of Helicobacter pylori infection in children does not necessarily result in persistent infection in all cases [44, 45].
We found a significant gender difference in the prevalence of Helicobacter pylori colonization, boys being infected more often than girls. These findings can not be confounded by age differences between the sexes, as the boys were younger than the girls on average but had a higher overall prevalence of Helicobacter pylori. These findings are comparable with findings in adults and also a study in children from Cameroon  but no such gender difference could not be found in a meta-analysis of 10 studies conducted over the last 20 years .
Congested living with more than 4 people in the household was associated with an increase in Helicobacter pylori prevalence, and this is similar to findings in other studies .
In our study almost all of the participants used a pit latrine and shared toilet with other household. Not sharing the toilet with other families was a risk factor. The strong association with 'sharing toilet' could be a spurious association. Not including this variable in the analysis did not alter the adjusted model. We are uncertain about the interpretation of this association.
In our study the prevalence of Helicobacter pylori colonization was not increased with decreasing socio-economic status of the family as found in other studies from the same region [7, 10, 47]. A possible explanation is the small socio-economic differences in our study population as our study was conducted in one parish only and within an area characterized by informal settlements, congested living, lack of proper sanitation condition, low education level among adults and small variation in income per family.
Few of the Helicobacter pylori colonized children complained about abdominal pain and only 12 children, with a mean age (± SD) of 6.6 (3.3) years were treated with an eradication cure. However the role of Helicobacter pylori in children with recurrent abdominal pain is controversial [48–51]. This study did not support a relation between abdominal pain and colonization with Helicobacter pylori.