Campylobacter pylori infection in biopsy specimens of gastric antrum: laboratory diagnosis and estimation of sampling error.

Campylobacter pylori infection was sought in 382 consecutive patients referred for upper gastrointestinal endoscopy. Five antral biopsy specimens were taken from each patient: one was inserted into a CLO-test to detect the urease activity of C pylori, two were sent for histological analysis where multiple sections were stained by the Warthin-Starry silver method, and two were sent for microbiological evaluation by Gram stain and culture. A patient was deemed to be infected when C pylori was cultured or seen in either the histological sections or the Gram stain of the biopsy smear. One hundred and seventy four (46%) patients were infected. Culture, Gram stain, histological examination and the CLO-test showed sensitivities of 92%, 87%, 93% and 90%, respectively. In 27 (15%) infected patients an uneven distribution of C pylori was seen between samples in the biopsy pair sent for histology. Examination of multiple sections stained with Warthin-Starry silver was more sensitive at detecting infection (93%) than examination of multiple sections from only one biopsy specimen (84%). Fifty seven of 80 patients, biopsied a median seven days (range 5 to 55) after completing colloidal bismuth subcitrate treatment, were still infected with C pylori. There was no decrease in the sensitivities of the above tests to detect infection after treatment. It is concluded that at least two antral biopsy specimens should be examined when attempting to diagnose C pylori infection by histological methods.