Editor, —We read with interest the papers by Reillyet al (Gut 1997; 40: 454–8) and by Christie et al (Gut 1996; 39:27–30) focusing on the Helisal saliva assay (Cortecs Diagnostics, UK) for the diagnosis of Helicobacter pyloriinfection. These papers reported that this test has a sensitivity of 84–88% and a specificity of 70–71%.

Editor,—We read with interest the papers by Reilly et al (Gut 1997; 40:454–8) and Christie et al (Gut 1996;39: 27–30) focusing on the Helisal saliva assay (Cortecs Diagnostics, UK) for the diagnosis of Helicobacter pyloriinfection. These papers reported that this test has a sensitivity of 84–88% and a specificity of 70–71%.

Firstly, in both studies the inter- and intra-assay coefficients of variation were not measured. This may well affect the reproducibility of a test. In fact, Fallone et al,1 using the Helisal assay, reported a sensitivity of 66%, a specificity of 74% and a coefficient of variation of 16±12%. Simor et alreported a sensitivity of 81% and a specificity of 75%, but the reproducibility of the Helisal assay results was not evaluated.2

Secondly, it is worth noting that in the study by Christie et al the age range of the reference population is not given. As Reilly et al found in their series, the Helisal saliva assay performs differently according to age (that is, 91.7% sensitivity and 35.3% specificity in the subgroup of patients under 50 years of age). Furthermore, Christie et al suggest that the Helisal saliva assay may be useful in reaching a diagnosis in children but it is conceivable that, as with H pyloriserology,3 salivary antibody testing for H pylori needs to be validated separately in adults and children.

Thirdly, Reilly et al comment that the Helisal saliva assay they used, when tested clinically by Lahaie et al,4 has a sensitivity of 97% and a specificity of 90%. However, in the latter study, published as an abstract, the figures refer to serum and not to saliva: 93% sensitivity and 85% specificity are the correct figures. Furthermore, these authors evaluated a relatively small number of patients, used a¹⁴C-urea breath test as the absolute reference, and did not give sufficient methodological details.

Finally, we believe that salivary H pylori testing may play an important role in selecting candidates for vaccination programmes.5 However, optimal validation of any diagnostic salivary kit for H pylori is warranted in order to assess reproducibility and whether it can be applied generally to both adults and children.6-8