The accuracy of serologic diagnosis of Helicobacter pylori infection in school-aged children of mixed ethnicity.

Abstract

The present study evaluated two non-invasive diagnostic methods for H. pylori infection in children, i.e. an in-house ELISA using sonicated Campylobacter jejuni antigen for absorption of cross-reacting antibodies and an immunoblot kit (Helico Blot 2.0, Genelabs, Singapore). 13C -Urea breath test (13C-UBT) was used as reference Sera and questionnaires were collected from 695/858 (81%) Swedish school children with mixed ethnic backgrounds within a cross-sectional, community-based study. Of 133 children with an ELISA OD value of > or = 0.1, all were screened with immunoblot and 107 made a 13C-UBT. The negative controls were 34/37 children from three school classes with an ELISA OD value of < 0.1 and volunteering for a 13C-UBT. An adjusted cut-off level for the ELISA of OD value 0.22 resulted in a sensitivity of 97.8%, a specificity of 95.8% and a concordance index of 97.2%. The Helico Blot 2.0 had a sensitivity of 97.8%, a specificity of 93.8% and a concordance index of 96.5%. The best concordance was seen for the 26.5 kDa (98.6%), 30 kDa (95.7%) and 19.5 kDa (91.5%) antigens. The corresponding concordance index for CagA was 78%, for VacA 73.8% and for the 35kDa antigen 68.8%. A significant difference in the distribution of the 19.5 and 26.5 kDa bands but not of CagA/VacA was noted by ethnic background. With an adjusted cut-off level for the enzyme-linked immunosorbent assay (ELISA), both non-invasive methods were found to have an adequate performance in a pediatric population. The differences in antibody response patterns by ethnic background represent a caveat in the interpretation of serological studies.