H. pylori: invasive and non-invasive test could diagnose it, but few could safely rule it out

Clinical bottom line (level 4)

  1. Two-thirds of patients with peptic ulcers or dyspepsia had H. pylori infections.
  2. Invasive and noninvasive tests could diagnose H. pylori infection.
  3. No chronic antral inflammation made H. pylori infection less likely. No other test could safely rule it out.
Cutler et al: Gastroenterology 1995; 109: 136-141
Expires November 2002

The study

Setting: gastroenterology clinic, acute hospital, USA

268 patients (aged mean 54, 53% male) with peptic ulcers (79%) or nonulcer dyspepsia, or undergone upper GI endoscopy for other reasons

Excluded if
  • previously received antibiotic therapy for H. pylori infection

Non-independent unblinded reference standard, applied in all patients from a consecutive appropriate spectrum.
Reference standard:

  • 4 tests or more concordant
Diagnostic test:
  • chronic antral inflammation
  • acute antral inflammation
  • Warthin-Starry stain
  • CLO test
  • urease breath test: measure 60 minutes after administration of 13-C urease
  • serum IgG: positive if optical density 1.0 or more
  • serum IgA: positive if optical density ratio 1.0 or more
 

The evidence

pre-test probability of H. pylori infection: 65%, (95% CI: 59% to 70%)

diagnostic test H. pylori infection no H. pylori infection LR+
(95% CI)		 post-test probability LR-
(95% CI)		 post-test probability
chronic antral inflammation 173 32 3.0
(2.2 to 3.9) 		84% 0.0
(0.0 to 0.026) 		0%
acute antral inflammation 150 6 14
(6.3 to 30) 		96% 0.14
(0.097 to 0.21) 		21%
Warthin-Starry stain 161 1 88
(13 to 620) 		99% 0.070
(0.041 to 0.12) 		11%
CLO test 155 0 -
(29 to infinity) 		100% 0.10
(0.067 to 0.16) 		16%
urease breath test 156 4 21
(8.2 to 56) 		98% 0.10
(0.065 to 0.16) 		16%
serum IgG 158 8 11
(5.6 to 21) 		95% 0.095
(0.058 to 0.15) 		15%
serum IgA 123 14 4.8
(3.0 to 7.9) 		90% 0.34
(0.26 to 0.43) 		38%
total 173 95

Comments

  1. Failure to provide an objective or blinded reference standard opens the study up to bias, and means the test characteristics will be overestimated.
  2. Practically these tests are sensitive enough to be reliable diagnostic tools for everyday use but we should keep in mind the sensitivity decreasing effect of antisecretory (proton pump inhibitor), antibiotic and bismuth therapy on these tests except for serology.

    Citation

    1. Cutler AF, Havstad S, Ma CK, et al: accuracy of invasive and noninvasive tests to diagnose Helicobacter pylori infection. Gastroenterology 1995; 109: 136-141
    Search Terms: pylori and diagnos* in Cochrane
    Contributor: Chris Ball and Clare Wotton, November 1999
    Reviewer: Martin Dawes

    Clinical Question.
    Patient peptic ulcer or nonulcer dyspepsia
    Intervention or Exposure histology, CLO test, urease breath test, serum IgA and IgG
    Outcome H. pylori infection