Determination of the test performance of less costly methods of <cite>Helicobacter pylori</cite> detection

Determination of the test performance of less costly methods of Helicobacter pylori detection

Carlo A. Fallone
Andrew Mitchell
William G. Paterson

Division of Gastroenterology, Hôtel Dieu Hospital, Gastrointestinal Diseases Research Unit, Queen's University, Kingston, Ontario

(Original manuscript submitted 15/9/94; received in revised form 3/1/95; accepted 3/2/95)

Abstract

Since Helicobacter pylori (Hp) has become so important in the management of peptic ulcers, it is imperative that noninvasive and inexpensive methods of diagnosis be available. The objective of this prospective, double-blind study was to determine the performance of the noninvasive 14C-urea breath test and the less expansive hematoxylin-phloxin-saffron (HPS) and Giemsa stains in the detection of this infection. As well, we set out to determine the minimal number of biopsies required to make an accurate diagnosis. Fifty consenting adults referred for gastroscopy underwent antral biopsies for HPS, Giemsa and Steiner silver staining and a 14C-urea breath test. The ability of HPS, Giemsa, and the breath test to differentiate between the presence or absence of Hp infection as defined by the silver stain was assessed using contingency table analysis. Hp was detected in 24 patients with the silver stain. Half of these patients had evidence of ulcer disease, compared to 3 of 26 patients who were negative for Hp on silver stain (p < 0.01). The Hp positive group were more frequently male (62% vs. 31%, p < 0.05) and more often had chronic active gastritis (96% vs. 7.7%, p < 0.001). The sensitivity of the breath test, HPS, and Giemsa stains were 95.8% (95% confidence interval: 79-100%), 75% (53-90%), and 95.8% (79-100%), respectively. Specificity was 100% (87-100%) for all 3 methods of detection. The breath test was also able to discriminate between heavily infected and moderately infected patients as defined by silver stain (p < 0.05), and could be shortened from 30 to 15 min without any loss of accuracy. Two antral biopsies were sufficient to make an accurate diagnosis. In conclusion, the 14C-urea breath test and Giemsa stain have excellent test performance scores, and can be used with confidence in place of the expensive silver stain. Hematoxylin-phloxin-saffron (HPS) also has excellent specificity but is inadequate to rule out Hp infection. Cost and discomfort could also be reduced by decreasing the number of antral biopsies to two, and by shortening the breath test from 30 to 15 min.

Clin Invest Med 1995; 18 (3): 177-185

Table of contents: CIM vol. 18, no. 3