Problem of distinguishing false-positive tests from acute or transient Helicobacter pylori infections.

Background: Reliable detection of acute Helicobacter pylori infections remains problematic. The high prevalence of false-positive non-invasive tests in low H. pylori prevalence populations makes identification of acute and transient infections difficult.

Methods: We explored the use of serum pepsinogens (PG) for diagnosis of acute infection in patients following H. pylori challenge such that the onset of the infection was known. We then compared those findings to a group of children with presumed acute infections defined as a positive urea breath test (UBT) and negative IgG serology.

Results: We examined the pattern and calculated cut-off values of PG levels in 18 adult volunteers with known acute H. pylori infection. We then compared the results with sera from nine symptomatic children with presumed acute H. pylori infection and a matched control group of nine children who did not meet criteria for acute H. pylori infection. In acute infection, both PGI and II levels increased following H. pylori infection reaching a peak by 2 weeks post-infection. The frequency of a positive test defined as a value > mean +2 SD was 17, 71, and 94% at week 1, 2, and 4 post-infection, respectively. Only one child with presumed acute H. pylori infection had an elevated serum PGI and one had an elevated PGII. Five of the children had follow-up UBTs and four were negative consistent with the diagnosis of false-positive UBT. H. pylori infection was confirmed in the child with an elevated PGI level.

Conclusions: These data suggest that a single positive noninvasive test in populations of low prevalence is most likely a false-positive result. This suggests that a single positive test requires confirmation preferably using a test that measures a different parameter (e.g., UBT confirmed by stool antigen test). It appears that most "transient"H. pylori infections are diagnosed on the basis of false-positive tests. PG levels are possible candidates as the confirmatory test.