Is the “Pragmatic Approach” for H. pylori Diagnosis Effective?
Amir Samani, Stefano Serra, Eva M Szentgyorgyi, Rajkumar Vajpeyi, Hala El-Zimaity. University Health Network, Toronto, Canada
Background: Controversy exists over the most cost-efficient approach for Helicobacter pylori diagnosis. Since 2004, to reduce costs, our group routinely screened gastric biopsies with sections stained only with hematoxylin and eosin (H&E). Special stains (Giemsa, Warthin Starry, or immunohistochemistry) were performed at the discretion of the pathologist. The goal of this study was to evaluate the “pragmatic approach” in a surgical pathology setting. In addition, we sought to evaluate how having biopsies from only one site (either antrum or oxyntic) might affect the diagnostic accuracy.
Design: We randomly chose from our files biopsy specimens from 93 patients who had at least antral and oxyntic biopsies taken specifically for H. pylori diagnosis. All cases were interpreted during regular sign out following the “pragmatic approach”. We then stained each block with a silver stain and Helicobacter immunohistochemistry stain. Slides were coded, randomized, and reviewed independently by four pathologists and one trainee. Observers had no information on the clinical status of the patients. We used a visual analogue scale graded from 0 (absent) to 5 (maximum) to score H. pylori and inflammation on each slide. We did not consider any stain as the “gold standard” for identifying H. pylori. We considered a biopsy specimen positive when all observers agreed that bacteria with the characteristic morphologic features of H. pylori were present in at least one of the special stain slides.
Results: There were 28 H. pylori-positive and 65 H. pylori-negative patients. Following the “pragmatic approach” 3 of 28 (11%) H. pylori positive patients were interpreted as negative at regular sign out (sensitivity 89%). Most of these biopsy specimens had minimal chronic inflammation. Helicobacter were identified at both sites in 57% of patients, in the antrum only in 25%, and in the oxyntic mucosa only in 18%. Of the 65 H. pylori-negative patients one was interpreted as positive for the infection (specificity 98%).
Conclusions: When the “pragmatic approach” is used, pathologists tend to over- estimate the minimal amount of inflammation that can be seen with H. pylori infection. This suggests that pathologists need to lower their threshold of when to order special stains, or get special stains routinely. A second potential reason for under diagnosis of H. pylori could be when biopsies are only taken from one site. In this study, biopsies from the antrum or body only would have resulted in a significant false negative diagnosis. The “pragmatic approach” using H&E sections only, leads to inconsistent results..
Tuesday, March 1, 2011 9:30 AM
Poster Session III # 145, Tuesday Morning