Editor,—We read with interest the paper by Dominguez-Munos et al (Gut1997;40:459–62) describing an optimal test drink in the¹³C-urea breath test (¹³C UBT) for the diagnosis of Helicobacter pylori infection. In this study all H pylori negative subjects (adults with dyspeptic symptoms) had a negative result with the ¹³C UBT (specificity 100%) after different meals. In other studies, using¹³C UBT to document H pylori infection both in adults and children, the sensitivity of the test ranged from 92 to 100% whereas specificity was usually above 92%.1 ,2However, no explanation has been given for the occurrence of false positive tests. Methodological bias and problems in defining the cut off value are possible reasons. However, there are no explanations for some false positive tests.3 ,4 Here, we report two children with a positive ¹³C UBT resulting from the presence of urease positive bacteria other than H pyloriin the stomach.

A 14 month old girl operated on just after birth for a congenital diaphragmatic hernia and presenting with severe gastro-oesophageal reflux associated with oesophageal dilatation and swallowing dysfunction was referred because of gastro-oesophageal haemorrhage. Endoscopy revealed oesophageal dilatation, severe oesophagitis and gastric stasis. The gastric and duodenal mucosa appeared normal. She was treated for two months with H₂ receptor antagonists. Antral and fundal biopsy samples (n=5) showed mild gastritis and wereH pylori negative on histology (Giemsa staining). Direct examination and culture of gastric biopsy specimens were both negative for H pylori. Serum specific antibodies against H pylori (ELISA) were also negative. ¹³C UBT was abnormal (5.63 δ%_O; normal values <3 δ%_O). Culture of gastric secretions revealed gastric bacterial overgrowth with colonic bacteria known to have urease activity (that is, Proteus mirabilis).

An 8 year old boy operated on just after birth for gastroschisis was referred because of a six month history of abdominal pain. Physical examination was normal. Endoscopy revealed moderate gastric stasis. Examination and culture of both antral and fundic biopsy specimens (n=5) were negative for H pylori as were serum specific antibodies against H pylori (ELISA).¹³C UBT was slightly abnormal (3.25 δ%_O, normal values <3 δ%_O). Culture of gastric secretions revealed gastric bacterial overgrowth with species, including micrococcus, with urease activity.

These two cases demonstrate that hydrolysis of urea as a result of bacterial metabolism can occur in the stomach of H pylorinegative subjects, and that ¹³C-urea can be hydrolysed in the presence of urease from bacterial species other than H pylori. Several bacteria—for example, P mirabilis,Escherichia coli, Yersinia enterocolita, Klebsiella pneumoniae, Staphylococcus aureus, have urease activity, but they do not usually colonise the stomach. Gastric bacterial overgrowth was probably favoured by prolonged antisecretory treatment in the first case and by gastric emptying abnormalities in the second (intestinal malrotation associated with gastroschisis). Urease activity associated with H pylori infection usually causes greater excretion of¹³C than that observed in our two patients (5.6 and 3.25 δ%_O respectively). As the cut off value of 3.00 δ%_O has been validated in both adults and children2 ,3 and no technical bias occurred, false positive resuts can be ruled out in our patients.

In summary, the ¹³C UBT is a sensitive and specific method for the non-invasive detection of H pylori infection, but gastric bacterial overgrowth may lead to a false positive diagnosis. These patients may be wrongly considered to be H pylori positive if a single, non-invasive test is used. In some circumstances (long term use of antisecretory drugs or abnormalities of gastric motility) a low positive ¹³C UBT without other evidence of H pylori infection (serology, bacteriology, histology) may be suggestive of gastric bacterial overgrowth.Letters, Book reviews, Notes