Aim: To prospectively evaluate the validity of a rapid office-based diagnostic serological test (using capillary blood) in our population, taking as reference a combination of standard diagnostic methods, as well as to compare the results of this technique with those of "classic" serological tests (using venous blood).
Patients and methods: We prospectively studied 39 consecutive patients with symptoms of the upper digestive tract who had undergone oral gastroscopy. Gastric biopsies were taken for histology and rapid urease testing, and a 13C-urea breath test was performed. An enzyme-linked immunoassay that detects IgG antibodies against Helicobacter pylori was used as a "classic" serological test and the commercial kit FlexPack HP was used as a "rapid" serological test. The endoscopist, the pathologist and those responsible for reading the rapid urease test, the 13C-urea breath test and both serological tests did not know the results of the other diagnostic methods. Patients were considered H. pylori positive when at least two of the three validated tests (rapid urease test, histology, and 13C-urea breath test) revealed infection and were considered free of infection when all tests were negative.
Results: Thirty-nine patients were studied. Thirty-eight per cent were male (mean age 48 +/- 15 years). The prevalence of H. pylori infection detected by the gold standard was 69.2%. The sensitivity and specificity of the "classic" serological test was 96% 95% (CI: 79-99) and 91% (59-100). "Rapid" serological testing was positive in nine patients, negative in 28 and indeterminate in two. A single digital puncture was sufficient in 80% of the patients, 15% needed two and 5% needed three. Most patients (77%) had no preference for either type of serological test while 20.5% preferred digital puncture and 2.5% venous puncture. The sensitivity, specificity, positive predictive value and negative predictive value were 31% (16-50), 91% (59-100), 89% (52-100) and 36% (19-56) respectively. Sensitivity was unaffected by age but specificity was lower in patients older than 40 years (89% vs. 100%; McNemar's test: 8; p < 0.01). Kappa's coefficient between the "classic" and the "rapid" serological tests was 0.16 (SE 0.1) and McNemar's test was 12.2 (p < 0.001), which indicates that the prevalence of infection diagnosed by both methods was not homogeneous.
Conclusion: The "rapid" office-based serological test used in our study is of insufficient diagnostic accuracy to be used in clinical practice to identify H. pylori infection.