Evaluation of an IgG-ELISA-kit for diagnosing Helicobacter pylori associated gastroduodenal disease

Abstract

Determination of IgG-antibodies against H. pylori in dyspeptic patients may be used as diagnostic screening procedure in place of invasive endoscopy for the diagnosis of H. pylori-associated gastroduodenal disease especially in younger patients with a low risk of gastric cancer. Therefore commercially available ELISA kits, which provide the easiest and most convenient method, will become more important in the near future. We compared ELISA results using acid extracted proteins and lipopolysaccharide of H. pylori with a commercially available ELISA (Biometra, Göttingen, FRG). For this investigation we selected sera from 81 dyspeptic patients with known H. pylori status. These patients were classified into five categories according to their individual histological diagnosis and their serum anti- H. pylori antibody-levels, as determined previously: I, seronegative patients with normal mucosa and negative H. pylori status (n = 7); II, seropositive patients with normal mucosa and negative H. pylori status (n = 6); III, seronegative patients with gastritis or duodenitis (n = 23; 6 23 H. pylori-positive in the biopsy): IV, seropositive patients with gastroduodenal disease but with negative H. pylori status (n = 24); V, seropositive patients with H. pylori-associated gastroduodenal disease (n = 21). By using the cut-off point of the commercial test recommended by the producer corresponding results (both tests negative or positive) were obtained in 70 81 patients (86%). The agreement was 71% in class I, 83% in class II, III and IV, and 100% in class V. In nine cases of the 11 discrepant results the antibody-levels lay close to the cut-off points of the two test systems. The two remaining cases, which were highly seropositive in the commercial system but negative in our test were due to one dyspeptic patient with normal mucosa and one patient with H. pylori-positive gastritis. We conclude that both assays have comparable cut-off levels and do not differ significantly in sensitivity (our ELISA 95%) or specificity (our ELISA 79%).