The prevalence of Helicobacter pylori infection in gastric cancer.

The prevalence of Helicobacter pylori infection in gastric cancer patients is difficult to estimate because the infection may be lost from individuals with cancer or its precursor conditions. The bacterium does not colonize cancerous tissue and therefore studies using histopathology as a means of assessing infection require gastric biopsy material sampled separately from the tumour itself. Results from five such studies, with more than 50 patients, have demonstrated an H. pylori prevalence of 43%-78%. Studies using serological assessment of antibody presence have the advantage of enabling adequate case-control comparisons. Ten reported studies that were retrospective in design showed a range in prevalence from 52% to 89% among the cases and from 38% to 78% among the controls. Five studies showed a significant increase in infection prevalence among the cases, with odds ratios ranging from 1.6 to 4.2. The odds ratio increased in subjects who were younger. Retrospective studies may be biased in that the prevalence of H. pylori among cases may be systematically under-estimated. Three prospective serological studies showed a prevalence of 69%-95% among cases and 47%-76% among controls. All three studies had a significantly elevated odds ratio and a pooled estimate of this was 3.8 (95% CI: 2.3-6.2). In the pooled analysis there was a significant trend towards an increased odds ratio with increased time between the assessment of infection and cancer diagnosis. After 14 years of follow-up the overall infection prevalence among cases was 90% and the odds ratio was 8.7 (95% CI: 2.7-45.0). This might represent the best estimate of infection prevalence and associated risk.(ABSTRACT TRUNCATED AT 250 WORDS)