New noninvasive tests for Helicobacter pylori gastritis - Comparison with tissue-based gold standard

The current gold standard for diagnosing H. pylori gastritis requires antral biopsy for urease test (eg, CLOtest) and/or histology. We compared this gold standard to a new low-dose capsule-based 1 mu Ci[C-14]urea breath test (UBT) and a rapid serum test for anti-H. pylori antibodies (FlexSure HT) in 50 consecutive patients undergoing upper endoscopy. Antral biopsies within 3 cm of the pylorus were used for CLOtest and were stained with H&E and thiazine. Slides were reviewed by a single, blinded pathologist (Review) and compared to the pathology department report (Report). A true positive was defined as a positive CLOtest or Review. The prevalence of H. pylori infection by each test was: CLOtest 32%, Review 36%, gold standard 42%, UBT 38%, FlexSure HP 44%, Report 44%. UBT had sensitivity = 90%, specificity = 96%, positive predictive value = 95%, and negative predictive value = 93% and accurately determined H. pylori status in a subgroup of 10 patients who had completed treatment. The one false positive breath test occurred in a patient with 2+ chronic inflammation on biopsy and a positive antibody test. There were two false negative breath tests. Each had both negative CLOtests and negative antibody tests. FlexSure HP had sensitivity = 74%, specificity = 89%, positive predictive value = 88%, and negative predictive value = 77% in patients not previously treated, but had a high false-positive rate in the 10 patients after treatment. Pathology department Report and blinded pathologist Review had only 84% agreement, kappa = 0.67. Both CLOtest and UBT agreed better with Review (86%, 87%, kappa = 0.68, 0.73, respectively) than Report (78%, 77%, kappa = 0.53, 0.53, respectively) suggesting that Review is more accurate than Report. When biopsy is done, we recommend that a CLOtest be performed followed by histological examination if the CLOtest is negative to ensure adequate sensitivity. However, interpretation of histological slides has significant interobserver variability even in experienced hands and may be inaccurate at times. FlexSure HP has good PPV but inadequate sensitivity to rule out active infection and is not useful in patients after treatment. UBT is an excellent noninvasive test that should be considered in cases where EGD is otherwise not indicated.