Abstract
OBJECTIVES: To evaluate prospectively the validity of a new diagnostic method based on a saliva sample, taking as reference the breath test with 13C-marked urea, and to compare the results of this technique with another indirect method based on the detection of antibodies, "classical" serology using venous blood. METHODS: 48 individuals, 24 healthy volunteers and 24 consecutive patients with gastro-duodenal ulcer disease, were studied prospectively. Treatment during the previous month with gastro-erosive medication, antibiotics, proton-pump inhibitors or bismuth-derived drugs, prior treatment to eradicate H. pylori, gastric surgery and the presence of linked illnesses, were all considered criteria of exclusion from the study. For the diagnostic test in saliva a commercial enzyme-linked immunosorbent assay (ELISA, trademark Helisal) was used; and for blood serology, another commercial ELISA (Helico-G). The staff responsible for reading the saliva, serology and breath tests did not know the result of the other diagnostic methods. The result of the breath test with 13C-urea (TAU-kit) was taken as the reference standard for H. pylori infection. RESULTS: The mean age of the healthy volunteers was 23 +/- 0.7 years; and of ulcer patients, 55 +/- 18. The prevalence of H. pylori infection, valued by the gold standard, was 79.2% in the ulcer patients and 54% in the volunteers. The sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy of the saliva test in the ulcer patients were, respectively: 100% (95% CI, 79-99), 60% (17-93), 90% (68-98), 100% (31-97) and 92% (71-98). In the volunteers these figures were: 46% (20-74), 73% (39-93), 67% (31-91), 53% (27-78) and 58% (37-77). The serology results were better, with 100% sensitivity in both groups and outstanding diagnostic accuracy (92% and 96% for ulcer patients and volunteers, respectively). Concordance between serology and the saliva test in ulcer patients was perfect (kappa, 1). However, in asymptomatic individuals concordance was deficient (kappa, 0.28), and the prevalence of infection diagnosed with the two tests was not homogeneous (McNemar, 2.8; p < 0.05). CONCLUSION: The diagnostic test for H. pylori infection in saliva is lacking in diagnostic accuracy in healthy individuals, which indicates that it cannot be used for screening infection in the asymptomatic population. The technique is more valuable in ulcer patients, although it does not reach the specificity desirable. For these reasons, the saliva test evaluated in this study cannot be recommended for diagnosis of H. pylori infection.