FAILURE TO CONFIRM ASSOCIATION OF VAC-A GENE MOSAICISM WITH DUODENAL-ULCER DISEASE

Background: Mosaicism of the Helicobacter pylori vac A gene comprises two families of allelic variations of the signal sequence region (s1, s2) and of the mid-region (m1, m2), Initial studies suggested that pep tic ulcer disease correlated with the s1 subtype of vac A. We compared the prevalence of various vac A genotypes of H. pylori isolates obtai ned from duodenal ulcer (DU) patients and subjects with simple gastrit is. Those isolates with s1 type were further examined to determine whe ther the specific vac A s1 (s1a versus s1b) genotype enabled predictio n of gastroduodenal disease. Methods: H. pylori isolates were obtained from 38 patient with endoscopically documented DU and 39 individuals with asymptomatic H. pylori gastritis from Houston, Texas. The vac A g enotype of each isolate was determined by polymerase chain reaction (P CR) amplification of genomic DNA for specific regions of the vac A gen e. Those isolates with s1 vac A subtype were further examined to deter mine whether they had s1a or s1b mosaicism. Results: There was no diff erence in frequency of the s1 genotype of isolates obtained from patie nts with duodenal ulcer or asymptomatic H. pylori gastritis in this sa mple (84% versus 79%, respectively; P = 0.77). The s1/m1 vac A genotyp e was detected in isolates from 16 duodenal ulcer patients versus 15 w ith H. pylori gastritis (P = 0.82). Detailed analysis of the s1 region failed to show a correlation of either s1a or s1b with duodenal ulcer . Both s1a and s1b genotypes were detected in 24 strains, and both mi and m2 mid-gene PCR amplicons were seen in 16 strains. Conclusions: We were unable to use H. pylori vac A genotyping to predict type of gast roduodenal disease in our patient sample. This failure to confirm an a ssociation of vac A genotype and duodenal ulcer disease differs from s amples from other regions. This most likely represents an example of d ifferences in H. pylori strains infecting host populations in differen t geographic regions. This study confirms the importance of establishi ng statistical associations with isolates from widely separate geograp hic regions before concluding that disease-related associations exist.