Histologic assessment of Helicobacter pylori status after therapy: comparison of Giemsa, Diff-Quik, and Genta stains.

To compare the Giemsa, Diff-Quik, and Genta stains for detection of Helicobacter pylori in gastric biopsy specimens, we stained, coded, and randomized slides, which were then independently reviewed by two pathologists and one trainee. H. pylori was graded from 0 (absent) to 5 (maximal intensity). Negative cases were from recently cured patients to ensure that a background of chronic inflammation was present. The time required to complete the evaluation was tabulated. The pathologists interpreted 72 H. pylori-negative slides, of which 1 (1.3%), 2 (3%), and 3 (4%) were scored positive (each with the score of 1) with Diff-Quik, Genta, and Giemsa stains, respectively (P = ns). Of the 128 H. pylori-positive slides, 5 (4%) had false-negative results with Diff-Quik, 8 with Genta (6%), and 14 with Giemsa stains( 11%). No Grade 2 slides were missed by Genta or Diff-Quik stains, but 3 of 20 were missed by Giemsa stain. The combination of hematoxylin and eosin and Diff-Quik provides accuracy similar to that of the Genta stain but requires more processing time. No stain was excellent after therapy for detecting one or two bacteria per slide; this finding emphasizes the need for obtaining multiple biopsy specimen to exclude failure of H. pylori therapy.