記住我
Abstract
Background: The 13C-urea breath test (13C-UBT) is preferred for non-invasive detection of Helicobacter pylori (H. pylori); however, its accuracy drops when results fall between 2‰ and 6‰ (called the gray zone). This study aimed to evaluate the accuracy of 13C-UBT (cut-off point 4‰) between 2‰ and 6‰, find a more appropriate gray zone, and identify the factors influencing 13C-UBT.
Methods: Patients with 13C-UBT results 2‰–6‰, over an eight-year period, were studied. H. pylori infection was diagnosed if patients were positive for either Warthin–Starry staining or quantitative real-time polymerase chain reaction (real-time PCR), and excluded if both were negative. Accuracy of 13C-UBT under different cut-off points was calculated, and the factors affecting 13C-UBT were analyzed.
Results: A total of 208 patients were included, of whom 129 were H. pylori–positive. Sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of 13C-UBT were 71.32%, 83.54%, 64.08%, and 87.62%, respectively. When the cut-off point was changed to 2.15‰, the NPV of 13C-UBT reached a maximum (76.47%); when the cut-off point was changed to 4.95‰, PPV reached its maximum (93.22%). Therefore, the original gray zone (2‰–6‰) was adjusted to 2‰–4.95‰. Gastric antral intestinal metaplasia (OR = 3.055, 95% CI: 1.003–9.309) was an independent risk factor for false-negative 13C-UBT.
Conclusions: Accuracy of 13C-UBT over 2‰–6‰ was poor, and the gray zone was changed to 2‰–4.95‰. 13C-UBT results over 2‰–4.95‰ should be interpreted with caution during mass screening of H. pylori, especially for patients with gastric antral intestinal metaplasia.
Keywords: 13C-urea breath test, false negative, gray zone, helicobacter pylori, intestinal metaplasia
How to cite this article: Introduction 
Helicobacter pylori (H. pylori) is one of the major etiological factors for peptic ulcer disease, gastric cancer, and mucosa-associated lymphoid tissue lymphoma.[1] Also, the infection rate of H. pylori is over 50% both in eastern and western Asia.[2] The Kyoto global consensus report defines H. pylori-associated gastritis as an infectious disease and recommends eradication therapy for all H. pylori-infected individuals.[3] Several invasive and non-invasive diagnostic methods have been developed till date for the detection of H. pylori infection, with histology of gastric biopsy specimens being the gold standard.[4] However, endoscopic biopsy causes physical discomfort and is prone to sampling error since H. pylori tends to be heterogeneously distributed in the stomach.[5] Serology test is non-invasive and convenient although, it cannot distinguish between past and present infections.[6] The H. pylori stool antigen test can detect active H. pylori infection, but is less appropriate for mass screening since delayed delivery of stool samples may lead to the degradation of H. pylori antigens.[7]
The urea breath test (UBT) has been extensively used in clinical practice and mass screening[1] since it can accurately detect active H. pylori infection [Table 1], and is more convenient to use than the diagnostic methods mentioned above.[7]13C-UBT, using 13C labeled urea, is based on the potent urease activity of H. pylori in the gastric mucosa due to which it hydrolyzes 13C labeled urea into NH3 and 13CO2, and hence, its infection can be diagnosed via breath measurements. The results of 13C-UBT are eventually presented as delta over baseline (DOB) value. However, false-negative results may occur if the patient has used antibiotics, bismuth, or proton pump inhibitors (PPIs) four weeks prior to the test.[8],[9] In addition, the reliability of 13C-UBT is undermined in patients with active upper gastrointestinal hemorrhage, atrophic gastritis, intestinal metaplasia (IM), or partial gastrectomy history.[10],[11],[12],[13] Moreover, individual variations in DOB values due to different body masses cannot be neglected.[4] Therefore, instead of setting a strict cut-off point that is applicable in all circumstances, determination of a gray zone, in which the likelihood of both false-negative and false-positive results of 13C-UBT would be maximal, has been proposed to be more sensible.[14] Thus, for individuals with a DOB value within the gray zone, a second test or a different diagnostic method would be recommended to re-evaluate the H. pylori status.[14],[15] Currently, there are relatively few studies on the 13C-UBT gray zone, and the exact range of the gray zone remains a controversial issue, although it mostly lies between 2‰ and 6‰.[12],[15] Moreover, while the sensitivity and specificity of 13C-UBT have been proven to be greater than 95%,[16] studies assessing the accuracy of 13C-UBT within the gray zone are not yet available.
Table 1: Test performance for common non-invasive H. pylori diagnostic testsThis study aimed to evaluate the accuracy of 13C-UBT between 2‰ and 6‰, and to identify a more appropriate gray zone, if possible. Furthermore, the study explored the factors responsible for false-negative or false-positive results of 13C-UBT and provided some insights into the interpretation of 13C-UBT results in clinical practice.
Patients and Methods Study design and participants
This was a single-center observational study. All consecutive patients who received 13C-UBT at our center from June 2013 to January 2020 were screened based on their electronic medical records. Those who met the following inclusion criteria were included in the study: (1) having a DOB value between 2‰ and 6‰; (2) having undergone gastroscopy within six months of 13C-UBT; and (3) whose paraffin-embedded specimens were available for bacterial DNA extraction and quantitative real-time polymerase chain reaction (real-time PCR) to detect H. pylori infection. Patients were excluded from the study if they (1) received an eradication regimen or took antibiotics for more than three days between 13C-UBT and gastroscopy; (2) used PPIs, antibiotics, or bismuth four weeks prior to gastroscopy or 13C-UBT; (3) had a history of gastrectomy; and (4) had active upper gastrointestinal hemorrhage during gastroscopy. The study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Ethics Committee of our center (No. 243-01; 2021/5/25). Written informed consent was obtained from all patients included in the study.
Procedure of 13C-urea breath test
Patients received 13C-UBT following a standardized protocol under the guidance of a clinician. Briefly, patients were instructed to fast for at least 4 hours before the collection of a basic breath sample. A capsule containing 75 mg 13C-urea (Headway Bio-Sci Co., Ltd, Shenzhen, China) was administered, and another breath sample was collected 30 minutes later. The collected breath samples were analyzed using an isotope-selective, non-dispersive infrared spectrometer (Headway Bio-Sci Co., Ltd, Shenzhen, China); 4‰ was used as the cut-off point, as validated in previous studies.[15] DOB values <4‰ were considered negative, and those ≥4‰ were considered positive.
Endoscopic biopsy and histological evaluation
As usual, two biopsy specimens, one from the gastric antrum and the other from the gastric corpus, were obtained from patients undergoing endoscopic biopsy, according to our local guidelines.[20],[21] Thereafter, the specimens were paraffin-embedded and sliced, and Warthin–Starry (WS) staining was performed to detect H. pylori, and hematoxylin and eosin staining was performed to assess inflammation, atrophic gastritis, and IM according to the updated Sydney grading system.[22]
Real-time PCR for the detection of H. pylori in paraffin-embedded specimens
For real-time PCR detection of H. pylori infection, five 5-μm thick paraffin rolls were sliced from each gastric mucosal specimen, and DNA was extracted therefrom using a Bacteria Genomic DNA Kit (CoWin Biotech Co., Ltd, Jiangsu, China). The H. pylori–specific 23S rRNA gene was detected by real-time PCR using a Helicobacter pylori detection kit (CoWin Biotech Co., Ltd, Jiangsu, China). Positive and negative controls were used for each sample. Real-time PCR analysis was performed with an ABI 7500 instrument (Applied Biosystems, Foster City, CA, USA) using the following reaction conditions: an initial denaturation step at 95°C for 8 minutes, followed by 45 cycles of amplification at 95°C for 15 seconds and fluorescence collection at 60°C for 1 minute, and finally, cooling at 25°C for 1 minute. Details of the primers and probes used are presented in [Table 2]. The results were analyzed as follows: when the cycle threshold (Ct) value of 6-FAM (6-carboxyfluorescein) channel was >35, indicating that the initial concentration of DNA target was less than 1 genome per assay reaction, the test result was considered negative.[23] When the Ct value was ≤35, the test result was considered positive.
Table 2: Sequences of primers and probes for detection of 23 S rRNA gene of H. pyloriReference standard for the diagnosis of H. pylori infection
A patient was diagnosed with H. pylori infection when either WS staining or real-time PCR results was positive; diagnosis was negative when both results were negative.[4]
Statistical analysis
Statistical analyses were conducted using SPSS Statistics 20 software (IBM, Armonk, NY, USA). Continuous variables conforming to normal distribution were shown as mean ± standard deviation; otherwise, they were shown as medians (interquartile ranges). Categorical variables were shown as the number of cases (percentage). Sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) for the 13C-UBT, in the range of 2.0‰ to 6.0‰ with an interval of 0.1‰, were calculated. Inter-group comparisons of categorical variables were performed using the Chi-squared test or Fisher's exact test. Factors affecting the accuracy of 13C-UBT were analyzed using binary logistic regression. All P values were two-tailed, and a P value <0.05 was considered statistically significant.
Results Study population
A total of 53,987 patients received 13C-UBT at our center from June 2013 to January 2020; of them, 3,139 (5.81%) had a DOB value between 2‰ and 6‰. Among such patients, 208 (6.63%) fulfilled the inclusion and exclusion criteria [Figure 1]. Overall, 103 patients (49.52%) had DOB values ranging from 2‰ to <4‰ and 105 patients (50.48%) had DOB values ranging from 4‰ to ≤6‰. The mean patient age was 48.30 ± 17.12 years, and the median time between 13C-UBT and gastroscopy was 2 (IQR 3) months. Histologic evaluation revealed atrophic gastritis in 61 patients (29.33%) and IM in 71 patients (34.13%) [Table 3].
Figure 1: Flow chart of patient selection. 13C-UBT, 13C-urea breath test; DOB, delta over baseline; PCR, quantitative polymerase chain reaction; PPI, proton pump inhibitorDiagnostic accuracy of 13C-UBT between 2‰ and 6‰
According to our pre-defined criteria for H. pylori status, 129 patients (62.02%) were diagnosed with H. pylori infection, and 79 (37.98%) were not infected. Real-time PCR results, WS staining results, and the corresponding DOB values of the patients included are shown in [Table 4]. Of the 91 patients with negative WS staining, 12 (13.19%) were real-time PCR positive; of the 117 patients with positive WS staining, 63 (53.85%) were real-time PCR negative [Table 4].
Table 4: Real-time PCR results, WS staining results and the corresponding DOB values of the patients includedAmong the 103 patients with 13C-UBT results from 2‰ to <4‰, 37 were diagnosed with H. pylori infection. Of the 105 patients with 13C-UBT results ranging from 4‰ to ≤ 6‰, 13 were not infected. The sensitivity, specificity, NPV, and PPV of 13C-UBT between 2‰ and 6‰ were 71.32%, 83.54%, 64.08%, and 87.62%, respectively.
The false-negative and false-positive rates of 13C-UBT between 2‰ and 6‰ were 35.92% and 12.38%, respectively. Specifically, the false-negative rates in the 2‰–3‰ and 3‰–4‰ ranges were 28.79% and 48.65%, and the false-positive rates in the 4‰–5‰ and 5‰–6‰ ranges were 19.57% and 6.78%, respectively. Furthermore, the false-negative rate in the 3‰–4‰ range was significantly higher than that in the 2‰–3‰ range (P = 0.044). In addition, false-positive rate in the 4‰–5‰ range significantly exceeded that in the 5‰–6‰ range (P = 0.048).
A more appropriate gray zone for 13C-UBT
Using different values between 2‰ and 6‰ as the cut-off point, we found that the NPV of 13C-UBT reached a maximum of 76.47% (13/17) when the cut-off point was changed to 2.15‰, and the PPV of 13C-UBT reached a maximum of 93.22% (55/59) when the cut-off point was changed to 4.95‰ [Figure 2].
Figure 2: Negative predictive value (Δ) and positive predictive value (△) of 13C-UBT at various cut-off points. The negative predictive value reached the maximum of 76.47% at 2.15‰, and the positive predictive value reached the maximum of 93.22% at 4.95‰Exploring a more appropriate gray zone, we found the results of 13C-UBT from 2‰ to <2.15‰ to have an NPV of 76.47%, which was still lower than 90%, and was not significantly higher than the NPV from 2.15‰ to <4‰ (61.63%, P = 0.244). Therefore, the lower limit of 2‰ of the original gray zone should remain the same. The results of 13C-UBT from 4.95‰ to ≤6‰ had a PPV of 93.22%, which significantly exceeded the PPV from 4‰ to <4.95‰ (80.43%, P = 0.048). Therefore, the upper limit of 6‰ could be reduced to 4.95‰. Overall, the new gray zone was set as 2‰–4.95‰. The sensitivity, specificity, NPV, and PPV of 13C-UBT between 2‰ and 4.95‰ were 50.00%, 88.00%, 64.08%, and 80.43%, respectively.
Risk factors for false-negative or false-positive result of 13C-UBT
In 129 patients diagnosed with H. pylori infection, multivariate analysis showed gastric antral IM (OR = 3.055, 95% CI: 1.003–9.309, P = 0.049) as an independent risk factor for false-negative results of 13C-UBT [Table 5].
Table 5: Risk factors for false-negative or false-positive results of 13C-UBTAmong the 79 patients without H. pylori infection, neither atrophic gastritis nor IM was associated with false-positive results in 13C-UBT [Table 5].
Discussion The prevalence of dyspepsia in Saudi Arabia is the highest in the Gulf region.[24] Since the test-and-treat strategy with non-invasive test is cost-effective for the initial management of dyspepsia,[25] it is recommended in countries with a low gastric cancer rate.[1] Among all the non-invasive diagnostic tests, 13C-UBT is the best option owing to its excellent performances.[1] In a recent meta-analysis, the sensitivity and specificity of 13C-UBT were reported to be 97% and 96%, respectively.[26] Satisfactory accuracy of the test makes the definition of a gray zone seem redundant. However, most studies concerning 13C-UBT have been conducted with a sample size in hundreds.[15] As is well known, gray zone results only occur in a remarkably low percentage of patients (approximately 2%)[15] and thus, exert only little influence on the accuracy of 13C-UBT in such studies. China is the most populous country in the world, with approximately 55.8% of its people infected with H. pylori.[2] When screening for H. pylori with 13C-UBT, a considerable proportion of the results obviously fall in the gray zone; for example, in our study, 3,139 (5.81%) of 53,987 patients had a DOB value between 2‰ and 6‰. Therefore, introduction of a gray zone would enhance the performance of mass screening for H. pylori, over that with a single cut-off point, since the former would inform clinicians about the results that would require further confirmation compared to those that would not.
The gray zone of 13C-UBT was first proposed by Mion et al.,[14] over a range of 2.5‰–3.5‰. Till date, other gray zones, such as 2‰–5‰, 3.5‰–4‰, and 2.5‰–6‰ have also been proposed,[15],[12],[27] most being between 2‰ and 6‰. 13C-UBT results greater than 6‰ are basically true positives, and those less than 2‰ are largely true negatives.[15] Kwon et al.[12] found that although the sensitivity and specificity of 13C-UBT (cut-off point 4‰) were only 68.9% and 84.9%, respectively, in patients with a high prevalence of atrophic gastritis (28.7%), DOB values less than 2.5‰ had an NPV of 98.48% and those greater than 6‰ had a PPV of approximately 90%. Since inaccurate results of 13C-UBT are basically distributed from 2‰ to 6‰, a gray zone calculated with only these results would be more precise than that using a vast range of DOB values. Therefore, we precluded 13C-UBT results less than 2‰ or greater than 6‰ in this study, and found the appropriate gray zone at around 2‰–4.95‰. In our study, 1.10% (594/53987) of the 13C-UBT results were distributed between 4.95‰ and 6‰. Due to the adoption of the new gray zone, such patients could be freed from repeat testing, thereby saving both time and money. For patients with 13C-UBT results in the new gray zone, a second 13C-UBT would be preferred in young patients, whereas histological evaluation would be recommended in older adults, especially those with alarm symptoms.[1]
We found the NPV of 13C-UBT to be only 64.08% between 2‰ and 4.95‰, and the detection rate of IM to be as high as 34.13% (71/208). Further analysis showed the gastric antral IM to be an independent risk factor for false-negative 13C-UBT, which had also been shown in previous studies.[11],[28],[29] Gastric IM is a crucial step in gastric carcinogenesis, and usually develops 10 years after initiation of gastric atrophy.[12],[30] In this study, 48.72% (19/39) of patients with antral IM were diagnosed with antral atrophic gastritis. Therefore, false-negative results of 13C-UBT were not only due to IM, but also due to atrophic gastritis, which had sabotaged the living environment of H. pylori for decades. First of all, H. pylori mainly colonizes the gastric antral mucosa. Atrophic gastritis and IM prevent the colonization of H. pylori.[29] and decrease its bacterial density. Lower the colonization density of H. pylori, lower are the 13C-UBT results.[31] Secondly, atrophic gastritis and IM reduce the number of G-cells in the gastric antrum and, therefore, downgrade the secretion of gastrin into circulation.[32] As a result, the secretion of gastric acid is impaired, and the gastric pH level rises. High pH levels inhibit UreI protein, a H+-gated urea channel regulating cytoplasmic urease, which is essential for the survival and colonization of H. pylori, and reduces the urea hydrolysis rate,[33] which eventually leads to false-negative results in 13C-UBT. According to the explanations above, gastric corpus IM could also be a risk factor for false-negative 13C-UBT. However, we failed to conclude the same, since only 11 patients were diagnosed with gastric corpus IM and none of them had false-negative 13C-UBT results. After all, gastric corpus atrophic gastritis and IM predominantly appear in autoimmune gastritis, whereas H. pylori–associated atrophic gastritis and IM usually develop in the antrum.[34]
Moreover, we found WS staining to have a false-negative rate of 13.19% compared to real-time PCR. The latter is known to be more sensitive than histology, especially in patients with a relatively low bacterial load.[35] Among the 12 patients with false-negative WS staining, atrophic gastritis with or without IM was histologically diagnosed in 5 (41.67%) patients. Therefore, false-negative WS results may have been due to the reduced bacterial load caused by atrophic gastritis and IM.[29] Taken together, when using histology to detect H. pylori in such patients, multiple biopsies should be performed to improve sensitivity,[36] or PCR should be performed, if possible.
Although this was a single-center study, the patients in this study were screened from 53,987 out-patients over a period of 8 years. Their mean and median age were 48.3 and 47.0 years, and were confined to normal distribution (P = 0.06). We also found the detection rate of atrophic gastritis, with or without IM, to vary in the different age groups and tend to increase with older age (<20 years: 0.00% [0/6]; 20–39 years: 10.45% [7/67]; 40–59 years: 32.47% [25/77]; 60–79 years: 49.02% [25/51]; >79 years: 57.14% [4/7]), which was in line with previous studies conducted in the southwest part of China, Israel, and Latvia.[37],[38],[39] Therefore, the patients in this study represented a larger population, and the conclusions should be helpful in decision-making regarding other regions as well.
This study had several limitations. First, it was intrinsically limited by its retrospective design; for example, selection bias might exist. Among the 3,139 patients with DOB values between 2‰ and 6‰, we only included 208 patients with concurrent endoscopic biopsy results, although there was no significant difference in age (P = 0.269), sex (P = 0.250), or DOB value distribution (P = 0.555) across the 208 patients included and 2,931 patients excluded. Second, unlike fresh specimens, paraffin embedding might have an adverse effect on the extraction of specimen DNA,[40] thereby affecting real-time PCR detection. Finally, the sample size was relatively small. Therefore, similar prospective studies should be conducted in future to avoid such limitations.
In conclusion, although the accuracy of 13C-UBT between 2‰ and 6‰ was poor, the PPV was ideal between 4.95‰ and 6‰; therefore, the gray zone of 13C-UBT could be changed to 2‰–4.95‰. For patients with 13C-UBT results in the new gray zone, a second 13C-UBT would be recommended in case of young patients, whereas histologic evaluation would be recommended, along with PCR, if possible, for older adults. The possibility of false-negative results would increase in patients with gastric antral IM.
Acknowledgements
We thank all the participants in this study for their cooperation, as well as the pathologists and nurses in the Department of Gastroenterology of Peking University Third Hospital for their excellent technical assistance in our research.
Financial support and sponsorship
This work was supported by the special funds of Beijing Key Laboratory of Helicobacter pylori Infection and Upper Gastrointestinal Diseases (Y57405-20).
Conflicts of interest
There are no conflicts of interest.
References 
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/sjg.sjg_638_21
留言 (0)