Mucosal Brush Sampling Increases the Helicobacter pylori Detection Rate of the Rapid Urease Test: A Pilot Study

Abstract

Background/Aims

Accurate diagnosis of and timely eradication therapy for Helicobacter pylori are crucial for managing and preventing adverse clinical outcomes associated with H. pylori infection. H. pylori infection is typically diagnosed using endoscopic biopsy-based tests such as the rapid urease test (RUT). In this study, we investigated the usefulness of mucosal brush sampling for H. pylori detection using the RUT and culture.

Methods

Twenty patients with H. pylori infection underwent endoscopy, and specimens were collected from the greater curvature of the gastric corpus via both mucosal brush and biopsy sampling methods. Brushing was performed using a disposable cytology brush, and the brush specimen was used for the RUT (brush-RUT) and then for culture. Two biopsies were obtained for the RUT (biopsy-RUT) and culture. H. pylori detection rates using RUT and culture yields from brush and biopsy samples were compared.

Results

The H. pylori detection rate was 100% with the brush-RUT, whereas it was 75% with the biopsy-RUT. Notably, among patients taking acid-suppressive agents, the sensitivity of the biopsy-RUT decreased to 66.7%, whereas that of the brush-RUT remained at 100%. The biopsy-RUT yield was also associated with the delta over baseline value determined by the urea breath test. H. pylori was successfully isolated from all the biopsy specimens and 95% of the brush samples, and all H. pylori isolates were tested for antimicrobial susceptibility. No significant procedure-related adverse events occurred with either sampling method.

Conclusions

Mucosal brushing is a simple, effective, and highly sensitive diagnostic method for H. pylori infection. The mucosal brush method is a practical alternative to biopsy, expanding the diagnostic capabilities while minimizing invasiveness.

INTRODUCTION

Helicobacter pylori is a Gram-negative bacterium that colonizes the human stomach.¹ Although its prevalence is gradually declining, H. pylori infection remains among the most widespread chronic infections worldwide.² H. pylori infection is associated with the development of chronic gastritis, peptic ulcer disease, gastric mucosa-associated lymphoid tissue lymphoma, and non-cardia gastric adenocarcinoma.^3,4 Therefore, accurate diagnosis and timely eradication therapy are crucial for the prevention of clinical outcomes associated with H. pylori infection.

Several invasive and noninvasive tests are available to detect H. pylori infection.^4,5 Invasive methods are based on endoscopy and include histopathology, the rapid urease test (RUT), culture, and molecular assays. Currently, no single gold standard for the diagnosis of H. pylori infection exists, and two or more complementary tests are often used to improve accuracy. In routine practice, endoscopic biopsy-based tests are the cornerstones of H. pylori diagnosis. The RUT is particularly popular because it is simple, inexpensive, and yields quick results.^6,7 This test detects the urease activity of H. pylori by a color-change reaction when a gastric sample (biopsy tissue, mucus, or gastric juice) is applied to a urea-containing medium. However, false-negative results can occur with the recent use of acid-suppressive agents, bismuth compounds, and antibiotics. Additionally, the non-uniform distribution of H. pylori in the stomach and the low density of organisms in some patients, such as those with atrophic gastritis or intestinal metaplasia, may cause sampling errors and contribute to false negatives. Various strategies have been suggested to improve RUT sensitivity, such as obtaining biopsies from both the antrum and corpus or using larger biopsy forceps; nonetheless, these involve obtaining more tissue and can increase the bleeding risk in certain patients.^6,8-10

Because the efficacy of RUT depends on collecting an adequate amount of urease-producing bacteria, less invasive sampling techniques have been explored to increase bacterial yield.¹¹ A previous study demonstrated that scraping the gastric mucus with open biopsy forceps produced a bacteria-rich sample and achieved higher RUT sensitivity than did the conventional biopsy method.¹² More recently, investigators have developed gastric mucus sweeping methods using devices such as fabric sheets or cotton swabs to sample a broad area of the stomach. These swab-based techniques have shown superior sensitivity and accuracy for H. pylori detection compared to biopsy tissue for RUT.^13-15 They also have the advantage of being less invasive than a forceps biopsy, although they require the preparation of a special swab material in advance.

Endoscopic brush cytology has been used as an adjunct to biopsy in the diagnosis of upper gastrointestinal lesions.¹⁶ The brush technique is advantageous for collecting samples from a larger surface of the stomach than that of a single biopsy, and is considered a reliable method for detecting H. pylori.^16-19 Indeed, some studies have reported that brush cytology is more rapid and sensitive than histology for identifying H. pylori, suggesting it as a valid alternative diagnostic tool.¹⁷ Nevertheless, conventional brush cytology has limitations; suboptimal brushing may yield inadequate samples for cytologic examination,¹⁹ and even a satisfactory smear requires microscopic evaluation by a pathologist to confirm the presence of H. pylori.^17,20 We speculated that collecting the gastric mucus alone, in which most H. pylori organisms reside, might suffice for diagnosis without the need for tissue biopsy or cytological analysis. Therefore, we aimed to compare the usefulness of the mucosal brushing method using a standard disposable cytology brush with conventional biopsy samples for RUT and culture.

MATERIALS AND METHODS

1. Study population

In this pilot study, adult patients with confirmed H. pylori infections scheduled for upper endoscopy were prospectively recruited between January 2025 and May 2025. Consecutive patients who visited the outpatient clinic were screened, and 20 individuals with a positive ¹³C-urea breath test (UBT; UBiT tablets, 100 mg; Otsuka Pharm Co., Ltd., Tokyo, Japan) were included. The interpretation of the UBT results was based on delta over baseline (DOB), with a DOB of >2.5‰ considered positive. The exclusion criteria were a history of gastric surgery or any contraindication to biopsy. Recent or current use of acid-suppressive agents (proton pump inhibitors [PPIs], potassium-competitive acid blockers, and H₂-receptor antagonists), as well as a history of H. pylori eradication therapy, did not lead to exclusion. Demographic information, medical histories, and endoscopic findings were also documented. The study protocol was approved by the Institutional Review Board of Hallym University Chuncheon Sacred Heart Hospital (IRB number: 2025-01-007), and written informed consent was obtained from each participant. This study was conducted in accordance with the principles of the Declaration of Helsinki.

2. Endoscopic procedure and sample collection

Upper endoscopy and sample collection were performed by gastroenterologists (C.S.B. and E.J.G.). Endoscopic findings were noted and classified according to the Kyoto classification.²¹ After completing the diagnostic examination, the gastric mucosa of the corpus greater curvature was sampled using a disposable cytology brush (BC-24Q, Olympus, Tokyo, Japan) passed through the working channel of the endoscope. The brush was swept over the corpus mucosa by moving it back and forth five times across the mucosal surface. It was then retracted into a sterile sheath and withdrawn from the endoscope. The material collected on the brush was immediately used to inoculate the RUT kit (Helicosign Dry; GenBody Inc., Cheonan, Korea), defined as brush-RUT by gently touching the brush tip onto the test disc (Fig. 1). Subsequently, the brush portion of the device was cut off with scissors and placed in a sterile tube containing 1 mL of normal saline for bacterial culture. For comparison, two biopsies from the greater curvature of the corpus (avoiding the brushed area) were obtained using standard biopsy forceps. One biopsy specimen was placed in the RUT kit (biopsy-RUT), and the other was positioned in a sterile tube with saline for culture. All samples for culture were kept at 4℃ until transport to the laboratory (within 4 hours of collection).

Fig. 1.

Schematic of the study design. RUT, rapid urease test; MIC, minimum inhibitory concentration.

A positive RUT result was determined by a change in color from yellow to red within 60 minutes of sample placement in the kit at room temperature. If the RUT color remained yellow after 60 minutes, the result was considered negative. In a subset of 15 cases, the time to a positive color change was measured for brush- and biopsy-RUT.

3. Bacterial culture and antimicrobial susceptibility testing

For H. pylori isolation, each biopsy specimen was gently rubbed onto a selective Brucella agar (Difco; Becton Dickinson and Company, Sparks, MD, USA) plate supplemented with 7% defibrinated sheep blood (MBcell, Seoul, Korea) or heat-inactivated horse serum (Gibco, Auckland, New Zealand) and antibiotics (amphotericin B 5 μg/mL, vancomycin 10 μg/mL, trimethoprim 5 μg/mL, and polymyxin 2.5 IU/mL). The brush sample was cultured in two ways: the brush tip itself was rolled across the surface of an agar plate, and saline transport medium (containing mucus released from the brush after vortexing) was inoculated onto a separate plate. All plates were incubated at 37℃ in a microaerophilic atmosphere (10% CO₂) for 5 to 7 days. The suspected colonies were identified as H. pylori based on their typical morphology and positive biochemical tests (urease, catalase, and oxidase). For definitive identification, a polymerase chain reaction (PCR) targeting the glmM gene was performed on each isolate using primers (forward, 5' AAG CTT TTA GGG GTG TTA GGG GTT 3'; reverse, 5' AAG CTT ACT TTC TAA CAC TAA CGC 3') after DNA extraction (AccuPower PCR pre-mix kit; Bioneer, Daejeon, Korea).²² The confirmed H. pylori isolates were subcultured and stored at –80℃ in tryptic soy broth containing 15% glycerol for further analysis.

Antimicrobial susceptibility testing of H. pylori isolates was performed using the agar dilution technique with Mueller-Hinton agar (Difco) supplemented with 5% defibrinated sheep blood (MBcell) and 2-fold serial dilutions of each antibiotic. The antibiotics tested were amoxicillin, clarithromycin, metronidazole, tetracycline, and levofloxacin. The bacterial suspensions were prepared in phosphate-buffered saline to a turbidity equivalent of approximately 2.0 McFarland standard. A 5 μL drop of this bacterial suspension was inoculated onto each antibiotic-containing agar plate. The inoculated plates were incubated at 37℃ under microaerophilic conditions and examined after 72 hours. The minimum inhibitory concentration was defined as the lowest concentration that completely inhibited visible bacterial growth. The following resistance breakpoints were used to interpret results: >0.125 μg/mL, >0.25 μg/mL, >8 μg/mL, >1 μg/mL, and >1 μg/mL for amoxicillin, clarithromycin, metronidazole, tetracycline, and levofloxacin, respectively, according to the European Committee on Antimicrobial Susceptibility Testing guideline.²³ The isolates with minimum inhibitory concentrations above these thresholds were classified as resistant to the respective antibiotics.

4. Statistical analysis

Descriptive statistics were used to summarize the data. Categorical variables are presented as counts and percentages, and continuous variables as medians with ranges. Given the small sample size, formal comparative statistical tests were limited, and the analysis was primarily descriptive for this pilot investigation.

RESULTS

1. Characteristics of the study population

The baseline characteristics of the study population are summarized in Table 1. The median age of the 20 patients was 60 years (range, 36 to 74 years), and 13 (65%) were male. All patients were eradication-naive. The median DOB measured by UBT was 17.7‰ (range, 2.9‰ to 55.4‰). Regarding medications, three patients (15%) were taking antiplatelet agents (two and one on aspirin and clopidogrel, respectively), and 12 patients (60%) were actively using acid-suppressive agents at the time of diagnosis (seven, three, and two on PPIs, potassium-competitive acid blockers, and H₂-receptor antagonists, respectively). One patient also received fluoroquinolone (moxifloxacin) within 3 weeks before endoscopy.

Table 1.

Baseline Characteristics of the Study Population

Variable	Value (n=20)
Age, yr	60 (36–74)
Male sex	13 (65)
Previous eradication therapy	0
Endoscopic diagnosis
Chronic atrophic gastritis	11 (55)
Peptic ulcer disease	6 (30)
Gastric adenoma	2 (10)
Gastric cancer	1 (5)
Endoscopic findings
Open-type atrophy	20 (100)
Intestinal metaplasia	10 (50)
Sticky mucus	11 (55)
Diffuse redness	6 (30)
Spotty redness	19 (95)
Xanthoma	7 (35)
Mucosal swelling	7 (35)
Nodularity	1 (5)
Use of antithrombotic agents	3 (15)
Use of acid-suppressive agents	12 (60)

Data are presented as median (range) or number (%).

2. H. pylori detection using brush and biopsy samples

Regarding diagnostic outcomes, brush-RUT was more sensitive than the conventional biopsy-RUT (Fig. 2). Brush-RUT yielded positive results in all 20 patients with a sensitivity of 100%. Conversely, biopsy-RUT was positive in 15 of the 20 patients, with a sensitivity of 75%. In other words, five infections that were missed by biopsy-RUT were successfully detected by brush-RUT. Four of these five false-negative biopsy-RUT cases included patients taking acid-suppressive agents. Indeed, when analyzing subgroups, biopsy-RUT sensitivity decreased to 66.7% (8/12) in patients on acid-suppressive agents compared to 87.5% (7/8) in those not on such medications (Fig. 3). Notably, brush-RUT maintained 100% sensitivity in both groups, regardless of medication use. Moreover, the brush samples tended to produce a faster urease reaction: among 15 patients with recorded reaction time, the median time to a positive RUT color change was 1.0 minute (range, 0.17 to 17.0 minutes) for brush samples, compared to 3.0 minutes (range, 0.5 to 15.0 minutes) for biopsy samples. When the RUT results were analyzed according to DOB, biopsy-RUT detected only 50.0% of infections in patients with low DOB values (<10‰), whereas its sensitivity improved to 80.0% and 88.9% in those with intermediate (10‰ to <20‰) and high DOB values (≥20‰), respectively.

Fig. 2.

Results of the rapid urease tests (RUTs) and culture tests of mucosal brush and biopsy specimens.

Fig. 3.

Factors associated with rapid urease test (RUT) sensitivity. (A) RUT positivity according to the use of acid-suppressive agents. The yield of the biopsy-RUT decreased in patients treated with acid-suppressive agents, whereas the brush-RUT remained sensitive. (B) RUT positivity of biopsy samples according to the delta over baseline values of patients, which were determined by urea breath tests. RUT sensitivity was higher in groups with higher delta over baseline values.

H. pylori was successfully isolated from the biopsy tissues of all specimens. Moreover, brush-derived samples showed high culture success rates. Plating the brush tip directly onto the agar resulted in the growth of H. pylori in 18 of 20 cases (90%), and inoculating the transport saline after washing the brush in the medium yielded H. pylori in 18 of 20 cases (90%). By combining both brush culture methods, H. pylori was isolated from at least one brush sample of 19 patients (95%). One patient whose brush samples failed to grow H. pylori had advanced gastric cancer and was receiving PPI at the time.

All H. pylori isolates obtained from biopsy and brush samples were successfully tested for antimicrobial susceptibility. The antimicrobial resistance rates for amoxicillin, clarithromycin, metronidazole, tetracycline, and levofloxacin were 35%, 40%, 60%, 5%, and 45%, respectively. Five patients (25%) showed heteroresistance to amoxicillin (n=1), clarithromycin and metronidazole (n=1), amoxicillin and metronidazole (n=1), amoxicillin, metronidazole, and levofloxacin (n=1), and clarithromycin, metronidazole, and levofloxacin (n=1).

No significant procedure-related adverse events occurred with either sampling method. Brushing the gastric mucosa caused only minor scratches, and no bleeding was observed endoscopically. Conversely, biopsy caused oozing in most patients, although this was self-limiting in all but one case. In this exceptional case, the patient was taking clopidogrel and developed persistent oozing at the biopsy site, which required endoscopic hemostasis using hemoclips.

DISCUSSION

Here, we investigated the usefulness of mucosal brushing as a sampling method for H. pylori diagnosis and found it to be a promising, less invasive alternative to the conventional biopsy approach. By obtaining paired brush and biopsy samples from each patient, we directly compared their performances. Overall, the sensitivity of the brush-RUT was higher than that of the biopsy-RUT. Furthermore, H. pylori culture from brush-collected specimens was successful in most cases, suggesting that brush sampling of gastric mucus can provide an accurate diagnostic yield comparable to that of biopsy, while potentially reducing the invasiveness of the procedure.

RUT is the most widely used method for H. pylori diagnosis during endoscopy. However, RUT performance is affected by certain conditions. If a patient has recently been on PPIs or antibiotics, the bacterial density of H. pylori in the stomach may decrease, leading to false-negative RUT results.^24,25 The patch distribution of H. pylori in the stomach can also contribute to sampling error, especially in cases of atrophic gastritis or intestinal metaplasia.²⁶ To counteract these issues, studies have recommended taking multiple biopsy samples (from both the antrum and corpus) or using larger biopsy forceps, which can improve RUT sensitivity.^6,8-10,27 Nonetheless, these attempts may increase mucosal trauma and may not be feasible in patients at a higher risk of bleeding, such as those on antithrombotic agents or with coagulopathy.

Recently, alternative sampling methods that do not require tissue biopsies have been proposed for H. pylori detection. One approach is the gastric mucus sweeping method, which employs a fabric or cotton swab to thoroughly swab the gastric mucosa and collect mucus laden with bacteria.^13-15,28 Because this method does not remove the tissue as in a biopsy, it was suggested as a safe sampling method for RUT or PCR. Studies on the sweeping method have reported higher sensitivity and faster RUT positivity compared to those of conventional biopsy-RUT.¹³ Importantly, the performance of the sweeping method appears to remain robust in patients receiving PPI therapy. Even after H. pylori eradication therapy (when bacterial loads are low), the sweeping method yielded diagnostic results comparable to standard biopsy approaches.¹⁴ These outcomes are encouraging, although it should be noted that several false-positive cases using the sweeping method were documented.^13,14 Another study used cotton swabs and found that swab-based RUT had higher sensitivity and accuracy than did biopsy, with both methods exhibiting 100% specificity.¹⁵ In that study, quantitative analysis demonstrated that the bacterial load obtained by swabbing was up to 32 times greater than that from biopsy specimens, highlighting the benefit of collecting a broad mucosal surface sample.

Previous studies on brush cytology have primarily focused on microscopic detection or have used brushes in urea broth to detect urease activity.^17,20,29 These studies established that brushing is generally less invasive than forceps biopsy and, apart from minor superficial abrasions, brushing does not typically cause significant injury to the mucosa. Additionally, as for the sweeping method, because brushings involve a large surface of the mucosa where many organisms are present, mucosal brushings seem to be a more sensitive method, particularly when H. pylori density is relatively low. In the present study, 60% of the patients were actively taking acid-suppressive agents, a context that often diminishes the sensitivity of biopsy-based diagnosis. Remarkably, brush-RUT correctly identified all infected patients in this subset, whereas biopsy-RUT failed to detect H. pylori in one-third of the patients with acid suppression. We also observed a clear relationship between a lower bacterial load (as indicated by the DOB values on the UBT) and false-negative biopsy-RUT results, reinforcing the idea that conventional biopsies may miss infection when bacterial density is reduced. Another practical benefit of the brush method is that it can be safely performed without significant bleeding, even in patients taking antithrombotic agents. Taken together, these findings suggest that mucosal brushing is a practical and effective alternative that can reduce the risk of bleeding associated with biopsies and provide a more representative sample for H. pylori diagnosis.

In addition to improving RUT sensitivity, the brush method proved capable of obtaining specimens suitable for H. pylori culture. Culture is the most specific diagnostic method for H. pylori, and it offers the critical advantage of yielding bacterial isolates for antimicrobial susceptibility testing, as well as enabling further research on virulence determinants and pathophysiology. With the worldwide increase in H. pylori antimicrobial resistance and the consequent risk of eradication failure, determination of antimicrobial susceptibility profiles has become increasingly important for guiding eradication therapy. Unfortunately, routine culture of H. pylori is not widely available, as it is labor-intensive, time-consuming, and requires dedicated laboratories.³⁰ Previous reports have shown that H. pylori can be isolated from gastric brush samples, indicating that the bacterial load obtained by brushing can be sufficient for successful culture even when the organism density is sparse.^17,31 However, these studies used devices not routinely used in clinical practice.^17,31 In our study, we deliberately used a commercially available cytology brush to test whether culture would be feasible with no special equipment. In most cases, we could isolate H. pylori from brush-collected specimens, and the isolates obtained from brush samples allowed us to perform antimicrobial susceptibility testing. Enhancing the sensitivity of sample collection can theoretically increase the chances of false-positive results. Biochemical testing and PCR confirmed that the cultures from the brush samples were positive for H. pylori.

Recently, molecular tests have become available to detect H. pylori directly in gastric samples and to determine the genotypic resistance of H. pylori to certain antibiotics.^32,33 For patients in whom endoscopy is indicated, several approaches can be used for further analysis, and the sample embedded in the RUT kit can also be used for additional testing.^22,34,35 Researchers have also found that molecular tests can be applied to the gastric mucus present on biopsy forceps placed in water or RUT gel,¹² as well as to gastric juice samples collected during endoscopy.^33,36,37 Likewise, it is expected that mucosal brushing samples can also be used for further analysis, including high-throughput analysis. An additional observation from our study is that brush samples can be preserved for further analysis. We found that H. pylori could still be cultured from a brush sample that had been stored and frozen for several weeks (data not shown), suggesting that brush specimens could be banked for research or deferred testing if needed.

This study had some limitations. First, it included only patients with confirmed H. pylori infection and no uninfected controls were tested. Thus, we could not assess the specificity of the sampling methods or calculate their positive or negative predictive values in a mixed population. Second, the sample size was relatively small, reflecting the pilot nature of this study. Although the results are encouraging, a larger study is needed to validate the findings and obtain more precise estimates of diagnostic performance. Third, we limited sampling to the greater curvature of the corpus. Additionally, we did not systematically evaluate the optimal number of brushes required. It is unknown whether brushing other regions (such as the antrum or multiple sites) might further enhance the diagnostic performance. Notably, unlike biopsy forceps, which require separate specimens from each site, a single brush may sample multiple gastric locations sequentially, potentially enhancing the cost-effectiveness of the brushing method. However, it should be balanced against the procedure time and potential mucosal irritation. Finally, the use of a disposable cytology brush for H. pylori diagnosis introduces an additional cost compared to the standard forceps biopsy. Further cost-effectiveness analysis may help determine the practical benefits of the mucosal brushing method in routine clinical practice.

In conclusion, the mucosal brushing method using a cytology brush is a simple, practical, and reliable method for H. pylori diagnosis. It achieves high sensitivity in RUT, even in situations where conventional biopsy-RUT shows false-negative results, such as during PPI therapy. Brush sampling is less invasive than a biopsy and covers a larger mucosal surface area, which improves the detection rate and allows rapid confirmation. Importantly, brush-collected samples can be used not only for RUT but also for bacterial culture and antimicrobial susceptibility testing, expanding the diagnostic capabilities without requiring additional sampling. This study demonstrates that comprehensive testing for H. pylori infection, from rapid point-of-care diagnosis to culture and susceptibility analysis, can be accomplished using a single device. Further larger-scale studies will help to define the role of brush sampling in H. pylori diagnosis and treatment and potentially encourage the development of improved diagnostic sampling techniques for this common infection.