Current issues of ACP Journal Club are published in Annals of Internal Medicine


Diagnosis

Noninvasive tests were as accurate as invasive tests for detecting Helicobacter pylori

ACP J Club. 1996 Jan-Feb;124:16. doi:10.7326/ACPJC-1996-124-1-016


Source Citation

Cutler AF, Havstad S, Ma CK, et al. Accuracy of invasive and noninvasive tests to diagnose Helicobacter pylori infection. Gastroenterology. 1995 Jul;109:136-41. [PubMed ID: 7540995]


Abstract

Objective

To compare the diagnostic accuracy of 6 tests (3 invasive and 3 noninvasive) for the diagnosis of Helicobacter pylori.

Design

A blinded comparison of 6 independent tests (7 sets of readings) for detecting H. pylori.

Setting

A large urban U.S. gastroenterology clinic.

Patients

268 patients (mean age 54 y, 59% < 60 years of age, 53% men, 53% nonwhite races) who were not previously treated for H. pylori infection and who were referred for endoscopy.

Description of Tests and Diagnostic Standard

The presence of H. pylori was invasively confirmed by endoscopy evaluation of chronic and acute inflammation in the gastric antral biopsy specimens, by Warthin-Starry (silver) stain, and by the CLO test. Noninvasive tests included 13C-urea breath tests (UBT) and serologic examination for IgG and IgA antibodies for H. pylori. The diagnostic standard was concordance of ≥ 4 of the 7 test results.

Main Outcome Measures

Sensitivity, specificity, and likelihood ratios for a positive (LR+) and negative (LR+) test result.

Main Results

82 patients had duodenal ulcer, 49 had gastric ulcer, 8 had pyloric channel ulcers, 55 had nonulcer dyspepsia, and the remainder had noncontributory findings. 81% of patients had concordance of ≥ 6 of 7 test results. 65% of patients were considered to be H. pylori positive. The respective sensitivity, specificity, LR+, and LR- were 100%, 66.3%, {3.0, and 0}* for chronic inflammation; 86.7%, 93.7%, {13.8, and 0.1}* for acute inflammation; 93.1%, 99.0%, {93.1, and 0.07}* for Warthin-Starry (silver) stain; 89.6%, 100%, {approaching infinity, and 0.01}* for the CLO test; 90.2%, 95.8%, {21.5, and 0.01}* for UBT; 91.3%, 91.6%, {10.9, and 0.09}* for serum IgG; and 71.1%, 85.3%, {4.8, and 0.3}* for serum IgA. Warthin-Starry (silver) stain had the best combination of sensitivity and specificity but was not statistically superior to CLO, UBT, and IgG antibody tests in diagnostic accuracy (P = 0.18, P = 0.27, and P = 0.43, respectively).

Conclusion

The noninvasive 13C-urea breath test and serologic examination for IgG antibodies were as accurate as the invasive Warthin-Starry (silver) stain and CLO test in diagnosing Helicobacter pylori in untreated patients.

Source of funding: Not stated.

For article reprint: Dr. A.F. Cutler, Section of Gastroenterology, Sinai Hospital, 6767 West Outer Drive, Detroit, MI 48235, USA. FAX 313-493-6213.

*Numbers calculated from data in article.


Commentary

14C-urea breath test and Giemsa stain were sensitive for detecting Helicobacter pylori

With increasing media exposure, many patients know about H. pylori and now request testing. The studies by Cutler and Fallone and their colleagues expand our understanding of the diagnostic accuracy of H. pylori testing. Now we can address which tests are cost-effective and which patients should be tested.

To overcome the lack of a true "diagnostic standard," Cutler and colleagues reasonably used consistency of results among several tests to represent true infection. This group also examined a large population that included patients with gastroesophageal reflux disease and a lower pretest probability of H. pylori infection. This cohort may have produced less accurate positive and negative likelihood ratios than a more selected group of patients with peptic ulcer. Fallone and colleagues confirm that UBT and antral biopsies have essentially equivalent diagnostic accuracy.

The results from these studies indicate that CLO, Warthin-Starry (silver) stain, UBT, Giemsa stain, and tests for serum IgG have large positive likelihood ratios. Therefore, any positive H. pylori test result eliminates the need for further confirmatory testing. Because all of these tests have equivalent diagnostic accuracy in untreated patients, the method of testing should be chosen by cost and ease of use.

For patients with ulcers diagnosed by barium, H. pylori IgG testing probably will be the least expensive and easiest to obtain. For patients diagnosed at endoscopy, antral biopsies and a CLO test should be done. Only in the event of a negative CLO would the biopsy specimens be sent for special staining. Patients who do not have duodenal ulcers caused by nonsteroidal anti-inflammatory drugs, however, have a 95% to 99% pretest probability of infection. Empiric treatment in this subpopulation of patients with duodenal ulcers may be a reasonable alternative.

Current studies do not support the diagnosis or treatment of H. pylori infection in nonulcer dyspepsia (1). Future research may, however, define a subpopulation of patients with nonulcer dyspepsia for whom treatment of H. pylori infection would be advantageous. Physicians who choose to diagnose and treat on an individual basis should recognize that this currently represents the "art of medicine."

Documentation of H. pylori eradication is recommended only for patients with complicated or bleeding ulcers (1). In this important subgroup, eradication reduces the rate of recurrent bleeding ulcer (2). The rapid office-based serologic assays for H. pylori are less useful in post-treatment evaluation because of the delayed and variable IgG titer decrease. Laboratory-based serologic testing (enzyme-linked immunosorbent assay) requires that acute and post-treatment samples be run at the same time (3). UBT appears to be the ideal method to document eradication and seems to be diagnostically equivalent to, safer, and more cost-effective than the current standard of repeat endoscopy with antral biopsies. Lack of widespread availability of UBT may necessitate the continued use of endoscopy in some clinical settings.

Several additional caveats should be considered. A definitive protocol for UBT has not been established. Many different office-based serologic tests exist for H. pylori, with variable diagnostic accuracy, and may not approach the level reported by Cutler and colleagues. Additionally, all of the tests evaluated in these studies were done in untreated patients. A post-treatment group would have a lower pretest probability of infection. In this population, these same diagnostic tests would provide less information.

The rapidly expanding volume of H. pylori research will soon lead to clearer and more cost-effective guidelines. Watch for future developments.

Philip S. Schoenfeld, MD, MSEd
David J. Roberts, MD
National Naval Medical CenterBethesda, Maryland, USA


References

1. NIH Consensus Conference. Helicobacter pylori in peptic ulcer disease. JAMA. 1994;272:65-9.

2. Rokkas T, Karameris A, Mavrogeorgias A, et al. Eradication of H. pylori reduces possibility of rebleeding in peptic ulcer disease. Gastrointest Endos. 1995;41:1-4.

3. Cutler A, Schubert A, Schubert T. Role of H. pylori serology in evaluating treatment success. Dig Dis Sci. 1993;38:2262-6.