Introduction

It is estimated that half of the global population is infected with Helicobacter pylori (H. pylori), a gram-negative, spiral-shaped microaerophilic bacterium [1]. The bacterium has several unique chemical and physical properties which allow for successful infection in the very hostile environment of the gastric lumen [2]. Specifically, 4–6 polar flagella which allow for motility in the mucus layer of the stomach and production of urease which hydrolyzes gastric acid to ammonia are important pathogenic factors [3]. Chronic infection with H. pylori is a potent risk factor for gastritis, peptic ulcer disease, and gastric malignancies such as gastric adenocarcinoma and lymphoma [3]. For example, it is estimated that 36–47% of gastric adenocarcinomas worldwide are solely attributable to chronic infection with H. pylori [4]. The infection is typically acquired in childhood [5], and while the exact method of transmission is unclear, poor sanitary conditions and high density family households are known risk factors [6]. Highly effective therapy for H. pylori infection is available for most patients in whom a diagnosis has been made [7]. Accurate diagnosis and timely treatment of the infection improves outcomes for patients infected with H. pylori [8]. For example, eradication of H. pylori has been shown to reduce the risk of subsequent peptic ulcer disease and gastric adenocarcinoma, both of which can lead to significant morbidity and mortality [8, 9]. H. pylori infection can be diagnosed on biopsies obtained from esophagogastroduodenoscopy (EGD) using either special stains on histology [10], rapid urease testing [11], culture [12], or polymerase chain reaction testing [12], but performing an EGD in order to diagnose H. pylori infection is expensive, invasive, and not universally available. Certain noninvasive tests, such as the urea-breath test and stool antigen assays, are also limited by high cost and/or a long turn-around time to diagnose the infection [12]. In addition, most patients with H. pylori infection require more than one test during their diagnostic and therapeutic journey as confirmation of eradication after treatment for H. pylori is recommended [7]. The aim of this study was to determine the performance characteristics of two new stool-based enzyme-linked assays (EIA) tests designed to rapidly detect presence of H. pylori antigen.

Materials and methods

Study design, patients, and specimens

We prospectively recruited patients at five geographically diverse clinical sites from August 2017 to May 2018 in order to determine the diagnostic performance of a novel stool antigen tests in patients in the initial diagnosis of H. pylori infection. Three of the study sites were in North America (Minnesota, VA, USA), one in Europe (Essen, Germany), and one in Southeast Asia (Dhaka, Bangladesh). For the initial diagnosis claim, we included patients with symptoms of dyspepsia, gastritis, or peptic ulcer. All patients underwent esophagogastroduodenoscopy where at least six gastric biopsies were obtained for histological analysis. One additional gastric biopsy was obtained for rapid urease testing. All patients provided a stool sample. Exclusion criteria included asymptomatic patients, patients in whom the presence/absence of H. pylori was already known. Patients had to have refrained from antibiotics and bismuth compounds (e.g., Pepto-BismolTM) for 2 weeks prior to submitting a fecal sample. Proton-pump inhibitor (PPI) use was recorded if within 2 weeks of providing the fecal sample.

Stool test

All study sites collected stool specimens from patients within 48 h before or after the reference endoscopy procedure. Stool specimens were subsequently stored at 2–8 °C for up to 14 days then frozen if not tested. All stool specimens were analyzed at the central reference laboratory. Stool samples underwent testing with H. PYLORI QUIK CHEK™ test, a rapid membrane EIA (rapid EIA), and the H. PYLORI CHEK™ test, a microwell EIA (EIA) according to the manufacturer’s instructions (TechLab Inc., Blacksburg, VA). An optical density (OD) of ≥ 0.120 at single wavelength (450 nm) or ≥ 0.080 at dual wavelength (450/620 nm) were considered positive for the H. PYLORI CHEK™ test and visual results were used to determine a positive result for the H. PYLORI QUIK CHEK™ test. In addition, all stool samples were also tested using two commercially available enzyme immunoassays following the manufacturer’s instructions (Premier Platinum HpSA Plus; Meridian Bioscience (microwell) and ImmunoCard STAT! HpSA, Cincinnati, OH (rapid test)).

Histological testing

Gastric biopsies were stained with hematoxylin and eosin or modified Giemsa stain for routine histological analysis to determine the presence or absence of H. pylori.

Rapid urease test

A single gastric biopsy was placed in the gel of the CLOtest* Rapid Urease Test (Kimberly-Clark*) and then stored at ambient temperature. A positive test was defined as a change to the reference color within 24 h based on the manufacturer’s instructions.

Culture

H. pylori culture was not routinely performed at all study sites. However, results were included in a sub-analysis for completeness.

Definition of the composite reference method

Diagnostic performance of the stool antigen tests was determined by comparing results to a composite reference method (CRM, gold standard for H. pylori infections) that includes 3 possible tests performed on the biopsy: culture, histology, and rapid urease test results, where a positive diagnosis is made if 2 out of 3 tests are positive [13]. For the purposes of this study, the results from rapid urease testing and histology were utilized for case definition as these two tests were performed in all patients across all study sites (Table 1). In addition, an overall analysis which included indeterminate cases based on the CRM definition as well as data from sites which performed H. pylori culture based was also performed. A subgroup analysis which excluded patients in whom PPI exposure occurred within 2 weeks was performed (Table 1).

Table 1 Composite reference method for diagnosis of H. pylori infection from gastric biopsy using histology and rapid tissue urease
Full size table

Cross-reactivity testing

The specificity of the H. PYLORI QUIK CHEK™ and H. PYLORI CHEK™ assays was challenged by examining the reactivity of a wide range of common intestinal organisms and viruses (Tables 2 and 3). For the analysis, the bacteria were grown to early stationary phase (> 108 CFU/mL); McFarland Standard #4 and stock cultures of viruses were purchased. The cultures were diluted 1:10 in (i) fecal matrix that was negative for H. pylori (negative fecal pool) or (ii) fecal sample matrix that was spiked with H. pylori antigen (ATCC strain 43526) at 2–3 times the amount to produce a positive results (C95; positive fecal pool). The preparations were assayed in both assays and qualitative results were reported.

Table 2 Bacteria analyzed for cross-reactivity testing
Full size table
Table 3 Viruses analyzed for cross-reactivity testing
Full size table

Ethical considerations

All patients provided informed consent and the study protocol received internal review board approval at each individual study site.

Statistical analysis

The demographic results were compiled and descriptive analysis, including counts and percentages were performed using SAS Software (JMP Pro 14.1.0). The clinical sensitivity and specificity for the H. pylori antigen assays were determined by a comparison to the composite reference method by cross-classifying each case as clinically present or absent with 95% confidence interval [14, 15]. The analysis included continuity correction.

Results

Demographic data

Patient demographics and clinical data are presented in Table 4. Overall, 271 patients participated in the study across five distinct geographic locations with 223 (82%) of the collected stool samples being tested following a single freeze-thaw on the assays. The mean age was 46.2 years (range 19–82) and 68.2% of participants were female. The prevalence of H. pylori positivity based on the CRM ranged from 0 to 69.4% depending on the geographical location. At the North American and European test sites, the prevalence of H. pylori in among the participants was less than 10% whereas at the South-East Asian site, nearly 70% of participants were infected.

Table 4 Patient characteristics
Full size table

Diagnostic performance of the novel stool antigen test in comparison with the CRM

Detailed performance characteristics of the new stool antigen tests are presented in Tables 5, 6, and 7. Of the 271 patients, 10 patients had indeterminate results based on the CRM definition for this study, and thus, the cohort used for assay performance included 261 patients. The H. PYLORI CHEK™ assay had a sensitivity of 92% and specificity of 91% for the detection of H. pylori infection compared to the CRM. The H. PYLORI QUIK CHEK™ assay had a sensitivity of 91% and specificity of 100% for the detection of H. pylori infection compared to the CRM. Both tests had a negative predictive value of 97%. The positive predictive value for H. PYLORI CHEK™ was 76% and 98% for H. PYLORI QUIK CHEK™ (Table 5).

Table 5 Performance characteristics of H. pylori antigen assays
Full size table
Table 6 Performance characteristics of H. pylori antigen assays in patients without PPI use
Full size table
Table 7 Biopsy, culture, and rapid urease results compared with the H. pylori stool antigen testing
Full size table

Diagnostic performance of the novel stool antigen test in comparison with other commercially available immunoassays

In the N = 261 study population, the Premier Platinum HpSA Plus assay had a sensitivity of 87% and specificity of 87%. The Immunocard STAT had a sensitivity of 92% and a specificity of 97% (Table 5).

Diagnostic performance in patients with no PPI exposure

Among patients in whom no PPI exposure occurred (N = 182), the H. PYLORI CHEK™ assay had a sensitivity of 95% and specificity of 90% for the detection of H. pylori infection compared to the CRM. The H. PYLORI QUIK CHEK™ assay had a sensitivity of 93% and specificity of 99% for the detection of H. pylori infection compared to the CRM (Table 6).

Detailed comparison of histopathological, culture, and rapid urease results with stool assays

Table 7 shows detailed results of comparison of the endoscopic-based tests with the stool-based assays in the entire study cohort (N = 271).

Cross-reactivity

A total of 38 common intestinal bacteria (Table 2) and 6 common intestinal viruses (Table 3) were used to challenge the specificity of the H. PYLORI QUIK CHEK™ and H. PYLORI CHEK™ tests. No cross-reactivity was observed with all the negative results remaining negative and spiked positive samples remaining positive.

Discussion

The H. PYLORI QUIK CHEK™ and H. PYLORI CHEK™ tests are new stool antigen tests within this current study that showed excellent sensitivity and specificity for the diagnosis of H. pylori infection. The clinical performance of these assays was superior to the Premier Platinum HpSA Plus test that had a high false-positive rate lowering the positive predictive value to 68% in this study population. A highly specific and sensitive noninvasive test for H. pylori is clinically important as the high positive and negative predictive values aid in preventing misclassification and hence under or over treatment of this important condition.

In addition to aiding initial diagnosis, these rapid assays may offer a useful tool to assess response to therapy and predict noninvasively those patients requiring retreatment. H. pylori antigen in stool may be present for weeks following treatment and current guidelines recommend that stool antigen testing not be done until at least 4 weeks after completion of antibiotic therapy to confirm eradication [7].

The strengths of this study include the robust performance of the stool antigen test across multiple diverse geographic sites. Furthermore, the diagnostic performance of the new stool assays was compared with a rigorous CRM consisting of gastric biopsies in combination with histology and rapid urease testing. We also compared the new stool assays to other commercially available immunoassays to assess performance. Specificity of the new assays was challenged using gut bacteria and viruses. In addition, a subgroup analysis showed that the overall performance of this assay did not appear to be significantly impacted by recent PPI exposure, although further work is needed to determine whether the current recommendation to hold PPI use prior to testing for H. pylori is needed. Limitations of this study include the high, but not unexpected, variability in H. pylori positivity between the different study sites and the female predominant study population. We did not collect data on additional pathological findings on biopsies beyond H. pylori positivity. In addition, H. pylori culture was not performed at all study sites. Finally, our study criteria of avoiding bismouth compounds and antibiotics for 2 weeks prior to testing could theoretically have led to some false-negative tests; however, the impact of this is unlikely to change the overall conclusions about the diagnostic performance of the assays.

In conclusion, the H. PYLORI QUIK CHEK™ and H. PYLORI CHEK™ assays demonstrate excellent clinical performance compared the composite reference method. Potential advantages of these new assays include the high accuracy and rapid availability of results for initial diagnosis and assessing response to therapy. Further studies need to examine the impact of same day results on initiation of therapy and eradication rates in the treatment of Helicobacter pylori.