The diagnostic validity of the (13)c-urea breath test in the gastrectomized patients: single tertiary center retrospective cohort study.

BACKGROUND: This study was conducted to evaluate the diagnostic validity of the (13)C-urea breath test ((13)C-UBT) in the remnant stomach after partial gastrectomy for gastric cancer.
METHODS: The (13)C-UBT results after Helicobacter pylori eradication therapy was compared with the results of endoscopic biopsy-based methods in the patients who have received partial gastrectomy for the gastric cancer.
RESULTS: Among the gastrectomized patients who showed the positive (13)C-UBT results (≥ 2.5‰, n = 47) and negative (13)C-UBT results (< 2.5‰, n = 114) after H. pylori eradication, 26 patients (16.1%) and 4 patients (2.5%) were found to show false positive and false negative results based on biopsy-based methods, respectively. The sensitivity, specificity, false positive rate, and false negative rate for the cut-off value of 2.5‰ were 84.0%, 80.9%, 19.1%, and 16.0%, respectively. The positive and negative predictive values were 44.7% and 96.5%, respectively. In the multivariate analysis, two or more H. pylori eradication therapies (odds ratio = 3.248, 95% confidence interval= 1.088-9.695, P = 0.035) was associated with a false positive result of the (13)C-UBT.
CONCLUSIONS: After partial gastrectomy, a discordant result was shown in the positive (13)C-UBT results compared to the endoscopic biopsy methods for confirming the H. pylori status after eradication. Additional endoscopic biopsy-based H. pylori tests would be helpful to avoid unnecessary treatment for H. pylori eradication in these cases.

INTRODUCTION

Helicobacter pylori is known the primary cause leading to chronic atrophic gastritis and peptic ulcer disease.1 This bacteria also associated with gastric cancers and mucosa- associated lymphoid tissue lymphoma.2,3 World Health Organization has reported H. pylori as a group 1 carcinogen, and the eradication of H. pylori infection reduces the risk of gastric cancer development, and prevents recurrences of these diseases.4–6 For the treatment of gastric cancer, curative attempts to resect gastric cancer leave a gastric stump, creating the possibility of metachronous tumors arising from the remnant stomach. Thus, many post-surgical management strategies have been attempted to lessen the risk of gastric cancer recurrence, such as adjuvant chemotherapy, regular endoscopic surveillance, and H. pylori eradication. After surgical resection, gastric surgery places patients in a different condition from those with a previous normal stomach. Subtotal gastrectomy can decrease the gastric emptying time,7,8 and increased bile reflux cause the changes of the hydrophobic gastric mucosal barrier,9 or enhanced blood flow in the remnant gastric body.10 For these reasons, it is crucial to survive or to detect for H. pylori in the remnant stomach, and the reported rates of H. pylori infection in the remnant stomach after distal gastrectomy fall within a broad range (19%–70%).11–13

H. pylori infections can be diagnosed by a variety of invasive and non-invasive methods.14 Endoscopic biopsy is the gold standard, but it is invasive and is prone to sampling error because H. pylori tends to be heterogeneously distributed in the stomach.15 Serologic examinations are noninvasive and convenient, but do not accurately reflect infection status.16 The urea breath test (UBT) using 13C-labeled urea is a noninvasive test based on the potent urease activity of H. pylori in gastric mucosa and was developed to overcome the shortcomings of serologic testing. This test has been widely used because it has been reported to have a sensitivity and specificity greater than 90% for detecting H. pylori infection, and to be more convenient to use and safer for patients.17,18 For these reasons, the 13C-UBT is now routinely used for diagnosis of H. pylori infection. However, information on the diagnostic efficacy of the 13C-UBT for the detection of H. pylori in the remnant stomach after partial gastrectomy is limited.12,19–21 Furthermore, the diagnostic accuracy of this test is controversial for the reason that the changed gastric anatomy could affect the result of the 13C-UBT. In this present study, we evaluated the diagnostic validity of the 13C-UBT after H. pylori eradication by comparing 13C-UBT result with that of endoscopic biopsy in the gastrectomized patients.

MATERIALS AND METHODS

Patient selection

Between January 2005 and March 2014, among 386 patients underwent the 13C-UBT after partial gastrectomy (Billroth I,22 Billroth II,23 Roux-en-Y anastomosis,24 wedge resection and proximal gastrectomy25) for gastric cancer in Seoul National University Bundang hospital. We retrospectively enrolled 161 patients with a 13C-UBT result after H. pylori eradication and that subsequently underwent follow up endoscopic surveillance to confirm final H. pylori status. Three endoscopic biopsy methods (histology [the modified Giemsa test], the rapid urease test [CLOtest, Delta West, Bentley, Australia], and culture in anastomosis site and lesser curvature and grater curvature of body) were used, and the final H. pylori infection was determined by revealing H. pylori infection at least any two tests of three. Endoscopic surveillances were performed at least on two occasions during follow-up after 13C-UBT without any further H. pylori treatment. The exclusion criteria applied were: (1) no follow up history of endoscopic surveillance after 13C-UBT; (2) the administration of antibiotics or the consumption of bismuth salts within 4 weeks or the administration of a proton pump inhibitor (PPI) within 2 weeks prior to 13C-UBT; (3) H. pylori eradication failure because of poor compliance; (4) a history of total gastrectomy; (5) H. pylori re-infection, where H. pylori status became positive after more 1 year after successful eradication.26 The study protocol was approved by the Ethics Committee at Seoul National University Bundang Hospital (SNUBH B-1406).

Helicobacter pylori eradication

For the treatment of H. pylori infection, PPI-based triple therapy (standard dose of PPI b.i.d., clarithromycin 500 mg b.i.d., and amoxicillin 1 g b.i.d. for 1 week) was used as a first-line therapy in all study subjects.27 When these first-line therapies failed, two types of rescue therapies were used, that is, bismuth-containing quadruple therapy [PPI b.i.d., tripotassium dicitrate bismuthate 300 mg q.i.d. (three tablets 30 minutes before meals and one tablet 2 hours after dinner), metronidazole 500 mg t.i.d., and tetracycline 500 mg q.i.d.] for 1–2 weeks,28 or moxifloxacin-containing triple therapy (moxifloxacin 400 mg q.d., amoxicillin 1 g b.i.d., and PPI b.i.d.) for 1–2 weeks.29 When second-line therapy failed then the other rescue therapy was used.

13C-urea breath test

Before the 13C-UBT, patients were instructed to stop taking medications (such as bismuth salts or antibiotics for 4 weeks and PPI for 2 weeks), and fasted for a minimum of 4 hours. After washing the oral cavity by gargling, a predose breath sample was obtained. Then, 100 mg tablet of 13C-urea (UBiTkitTM, Otsuka Pharmaceutical Co. Ltd., Tokyo, Japan) was administered free of citric acid. Breath samples were collected in the sitting position using special breath collection bags before 13C-urea administration (baseline) and 20 minutes after administration. Collected breath samples were analyzed using an isotope-selective, nondispersive infrared spectrometer (UBiT-IR 300®, Otsuka Pharmaceutical Co. Ltd, Tokyo, Japan). Despite the lack of validation, a 13C-UBT cut-off value 2.5‰ was used as recommended by the manufacturer, and thus, a delta 13CO2 of ≥ 2.5‰ was considered positive for H. pylori infection.

Endoscopic surveillance for the detection of Helicobacter pylori infection and histological evaluation

To improve the H. pylori detection rate, three gastric biopsies were obtained from the antral side of the anastomotic site, lesser curvature and greater curvature of body in the remnant stomach. The presence of H. pylori was assessed by modified Giemsa staining and degrees of inflammatory cell infiltration, atrophy, and intestinal metaplasia were assessed by hematoxylin and eosin staining. Histological features of gastric mucosae were graded using the updated Sydney scoring system, which has a four-point scale (i.e., 0 = none, 1 = slight, 2 = moderate, and 3 = marked).30

Mucosa urease test

Two biopsy specimens, one from the lesser curvature of the anastomosis site and lesser curvature of body, were used for the rapid urease test (CLOtest, Delta West). Anastomosis site and body biopsy specimens in the remnant stomach were evaluated separately, and all urease tests were monitored for color change for up to 24 hours.

Microbiological examination

Two specimens from the antrum and body were sent for microbiological culture in brain heart infusion plates containing 7% horse blood. These plates were placed in a glass tank in a 5% O2, 10% CO2, and 85% N2 atmosphere at 37°C for 3–5 days. Anastomosis site and body biopsy specimens in the remnant stomach were evaluated separately. Organisms were identified as H. pylori by Gram staining, colony morphology, and positive oxidase, catalase, and urease reactions.

Statistical analysis

Sensitivity, specificity, false positive rate, false negative rate, positive predictive value, negative predictive value for the 13C-UBT in the partial gastrectomy state were calculated. Statistical analysis was conducted using PASW Statistics ver. 18.0 (IBM Co., Armonk, NY, USA). The Student’s t-test, Pearson’s chi-square test, and Fisher’s exact test were used, as appropriate, for the univariate analysis of factors affecting the accuracy of the 13C-UBT, and a logistic regression model was used for the multivariate analysis. P-values of < 0.05 were considered statistically significant.

RESULTS

Subject characteristics

Of the 161 patients with a 13C-UBT result after partial gastrectomy state, forty seven patients (29.2%) had a 13C-UBT result ≥ 2.5‰ and 114 patients (70.8%) had a value < 2.5‰. Baseline patient characteristics are shown in Table 1. There were 97 males (60.2%) and 64 females (39.8%) with a mean patient age of 58.6 (30–78) years. Indications for operations were early gastric cancer in 125 patients (77.6%) and advanced gastric cancer in 36 patients (22.4%). Analyzing method of gastrectomy, Billroth I, II anastomosis, Roux-en-Y anastomosis, wedge resection and proximal gastrectomy were carried out for reconstruction of the gastrointestinal tract in 106 (65.8%), 6 (3.8%), 29 (18.0%), 11 (6.8%) and 9 (5.6%) patients. 17 (10.6%) and 27 (16.8%) patients had diabetes mellitus and hypertension, respectively. Before the 13C-UBT, 110 patients (68.3%) underwent the first-line eradication therapy, 41 patients (25.5%) underwent second-line eradication therapy, and 10 patients (6.2%) underwent third-line eradication therapy (Table 1).

Table 1.

The baseline characteristics of patients with a 13C-UBT value

Characteristic	Results
Age	58.6 ± 10.9 (30–78)
Gender (Male:Female)	97 (60.2%):64 (39.8%)
Underlying disease
DM	17 (10.6%)
HTN	27 (16.8%)
Diagnosis for gastrectomy
Early gastric cancer	125 (77.6%)
Advanced gastric cancer	36 (22.4%)
Type of gastrectomy
Billroth I	106 (65.8%)
Billroth II	6 (3.8%)
Roux-en-Y anastomosis	29 (18.0%)
Wedge resection	11 (6.8%)
Proximal gastrectomy	9 (5.6%)
Total number of eradiation therapies for H. pylori infection
First	110 (68.3%)
Second	41 (25.5%)
Third	10 (6.2%)
Final UBT result
Positive	48 (29.8%)
Negative	113 (70.2%)

13C-UBT, 13C-urea breath test; DM, diabetes mellitus; HTN, hypertension; H. pylori, Helicobacter pylori.

The diagnostic accuracy of 13C-urea breath test in the gastrectomized patients

A flowchart of the study is shown in Figure 1. All 161 patients were evaluated for H. pylori status by histology and using the CLOtest. But H. pylori culture was performed in 8 patients during the endoscopic surveillance after the 13C-UBT. The success rate of culture was 87.5%. 7 patients were proved H. pylori infection on the culture methods. When H. pylori statuses was analyzed based on biopsy-based methods, 25 patients (15.5%) had a positive result for H. pylori infection, and 136 patients (84.5%) were H. pylori negative (Fig. 1).

Figure 1.

Flow chart of 13C-urea breath test (13C-UBT) and endoscopic biopsy based results of Helicobacter pylori (H. pylori) status after eradication in the partial gastrectomized patients. CLO, rapid urease test (Campylobacter-like organism).

In patients with a negative 13C-UBT result, histology and the CLOtest showed that 110 patients (96.5%) were H. pylori negative and 4 patient (3.5%) were H. pylori-positive. However, in the group with a positive 13C-UBT result (n = 47), histology, the CLOtest, and/or culture showed 21 patients (44.7%) were H. pylori positive and 26 patients (55.3%) were H. pylori negative (Fig. 1).

When the diagnostic accuracy of the 13C-UBT was calculated versus endoscopic biopsy results, its sensitivity, specificity, false positive rate, and false negative rate were 84.0%, 80.9%, 19.1%, and 16.0%, respectively. Its positive and negative predictive values were 44.7% and 96.5%, respectively (Table 2).

Table 2.

The sensitivity, specificity, and predictive values of the 13C-UBT compared with endoscopic biopsy methods for detecting the H. pylori status

		Using the biopsy-based methods for H. pylori status
		Positive	Negative
13C-UBT value	≥ 2.5‰	21 (13.1%)	26 (16.1%)	PPV 44.7%
	< 2.5‰	4 (2.5%)	110 (68.3%)	NPV 96.5%
		Sensitivity 84.0%	Specificity 80.9%

13C-UBT, 13C-urea breath test; H. pylori, Helicobacter pylori; PPV, positive predictive value; NPV, negative predictive value.

False positive results of the 13C-urea breath test in the partial gastrectomy state

Figure 2 shows a discordant H. pylori status results for the 13C-UBT and endoscopic biopsy in the ≥ 2.5‰ range (n = 26). The median 13C-UBT value of this discordant group was 5.9‰ (2.6–9 8‰). 73.1% (19/26 patients) in the 2.5‰ to 10.0‰ range, 26.9% (7/26 patients) in the > 10.0‰ range had a discordant result between 13C-UBT and endoscopic biopsy based results.

Figure 2.

Individual values of the false positive 13C-urea breath test (13C-UBT) results after Helicobacter pylori (H. pylori) eradication in the partial gastrectomized patients (median 13C-UBT value = 5.9‰).

The risk factors for a false positive result of the 13C-urea breath test after H. pylori eradication

The mean delta value of the 13C-UBT (≥ 2.5‰) after H. pylori eradication was not shown a statistical difference between matched (12.3‰) and unmatched group (13.5‰). Among those patients with a positive 13C-UBT result, 23 patients (48.9%), 20 patients (42.6%), and 4 patients (8.5%) underwent first, second, or third eradication therapies for H. pylori infection before enrollment in this study. Mean time from eradication to 13C-UBT was not significantly different for those with matching 13C-UBT and biopsy results. However, univariate analysis of the risk factor that caused mismatched results showed that total number of H. pylori eradication therapies (P < 0.005) significantly contributed to mismatching (Table 3). Logistic regression analysis confirmed that multiple prior eradication therapies (odds ratio = 3.248, 95% confidence interval = 1.088–9.695, P = 0.035) was associated with the mismatching of 13C-UBT and biopsy results after H. pylori eradication (Table 4).

Table 3.

Risk factors for the mismatched 13C-UBT positive result compared with biopsy based methods after H. pylori eradication in the partial gastrectomized patients

	13C-UBT result compared with biopsy based methods		Univariate P-value	Multivariate P-value
	Matched group (n = 21)	Mismatched group (n = 26)
Gender (Male:Female)	12 (57.1%):9 (42.9%)	17 (65.4%):9 (34.6%)	NA
Age (years)	59.7 ± 8.9	60.5 ± 10.7	NA
Underlying disease
DM	5 (23.8%)	2 (7.7%)	NA
HTN	5 (23.8%)	5 (19.2%)	NA
Type of gastrectomy			NA
Billroth I	10 (47.7%)	16 (61.5%)
Billroth II	1 (4.8%)	1 (3.9%)
Roux-en-Y	6 (28.6%)	4 (15.4%)
Wedge resection	4 (19.0%)	4 (15.4%)
Proximal gastrectomy	0 (0.0%)	1 (3.8%)
Time to 13C-UBT after eradication (weeks)	6.4	6.8	NA
The mean delta value of the 13C-UBT (‰)	12.3 ± 22.8	13.5 ± 15.1	NA
The total number of H. pylori eradication therapies			< 0.005	0.035
Single	16 (76.2%)	7 (26.9%)
Two	3 (14.3%)	17 (65.4%)
Three	2 (9.5%)	2 (7.7%)
The degree of gastric mucosal atrophy in the remnant stomach			NA
None	20 (95.2%)	20 (76.9%)
Mild	1 (4.8%)	4 (15.4%)
Moderate	0 (0.0%)	0 (0.0%)
Marked	0 (0.0%)	2 (7.7%)
The degree of gastric mucosal intestinal metaplasia in the remnant stomach			NA
None	19 (90.5%)	16 (61.5%)
Mild	0 (0.0%)	4 (15.4%)
Moderate	0 (0.0%)	3 (11.5%)
Severe	2 (9.5%)	3 (11.5%)

13C-UBT, 13C-urea breath test; NA, not available; DM, diabetes mellitus; HTN, hypertension; H. pylori, Helicobacter pylori.

Table 4.

Logistic regression model for risk factors of the mismatching result between 13C-UBT and endoscopic biopsy based methods in the diagnosis of H. pylori infection

Risk factors	β	S.E	P-value	OR	95% CI
Multiple eradication therapies for H. pylori infection	1.178	0.558	0.035	3.248	1.088–9.695

13C-UBT, 13C-urea breath test; H. pylori, Helicobacter pylori; β, coefficient; S.E, standard error; OR, odds ratio; CI, confidence interval.

DISCUSSION

The results of the present study suggest that there was a poor diagnostic concordance between the 13C-UBT and endoscopic biopsy-based results in the positive 13C-UBT values (≥ 2.5‰) after H. pylori eradication in remnant stomach after partial gastrectomy for gastric cancer. Generally, the 13C-UBT has been reported to be one of the most accurate diagnostic tools for assessing H. pylori status,31,32 and due to its speed, cost-effectiveness, and convenience, this test has been widely adopted in clinical practice.33 After partial gastrectomy, the gastric anatomy is altered and the test urea might be expected to pass through the stomach faster, giving different reactant percentages in reaction time, or the 13C-UBT results could be influenced by bile acid reflux. Several reports have shown that 13C-UBT provides lower diagnostic accuracy when using histology as a reference in the remnant stomach after partial gastrectomy.34–36 Previously, Lotterer et al.20 reported that the 13C-UBT had shown a 100% sensitivity and 80% specificity in patients with partial stomach resection. In a Japanese study, the sensitivity, specificity, and accuracy of the 13C-UBT were shown to be 96.3%, 100%, and 97.1% in the remnant stomach.21 However, Schilling et al.12 reported that the sensitivity of the 13C-UBT was 52%, the specificity 93%. The positive predictive value was 81.25%, the negative predictive value 76.9% and the accuracy was 77.9%. In our study, the sensitivity, specificity, false positive rate, and false negative rate of the 13C-UBT were 84.0%, 80.9%, 19.1%, and 16.0%, respectively. Its positive and negative predictive values were 44.7% and 96.5%. Overall, the diagnostic validity of the 13C-UBT was disappointing especially in the positive range of the 13C-UBT results.

Histological examination of the gastric mucosal biopsy specimen is generally considered to be the current gold standard for the diagnosis of H. pylori infection. In this study, the observed false positive 13C-UBT results raise the question as to whether endoscopic biopsy-based methods are reliable for determining the final status of H. pylori. A previous study showed that the 13C-UBT occasionally has poor diagnostic ability as compared with endoscopic biopsy-based methods.18 However, endoscopic biopsy-based methods are susceptible to sampling errors because of discontinuous H. pylori colonization of the stomach.37 In our study, gastric biopsies were performed and the samples were obtained from the anastomosis site and the body of the remnant stomach for histological analysis (modified Giemsa staining), and a gastric sample was obtained from both anastomosis site and body for rapid urease testing. H. pylori culture was performed using samples from the gastric biopsies, that is, 2 samples from the anastomosis site and from the body. Furthermore, all patients received at least two or more follow-up endoscopic surveillances, which re-confirmed final H. pylori statuses. Thus, we believe the possibility of a gastric biopsy sampling error was slight.

Considering a change of anatomical status of the stomach, postgastrectomy-induced hypochlorhydria often results in bacterial colonization of the remnant stomach. Urease-producing bacteria (Streptococcus, Staphylococcus, Gardnerella, Lactococcus, and Enterococcus) could cause false positive results.38–41 In a hypochlorhydric state of the remnant stomach, it accelerates the colonization and the overgrowth of non-H. pylori urease-positive bacteria. Furthermore, elevation of intragastric pH removes the neutralizing action of hydrochloric acid on local ammonia production by H. pylori urease, leading to the ultimate death of the bacterium as a result of overalkalization.42 Previous study showed that the persistence of H. pylori in the residual stomach decreased from 68.8% to 36% as time elapsed after surgery went from less than 1 year to more than 3 years without any eradication.19 According to a Taiwan study, spontaneous clearance of H. pylori develops in a certain number of patients who underwent distal gastrectomy.43 In this study, they reported decreased prevalence of H. pylori colonization was found after partial gastrectomy without additional eradication therapy: 1–15 years, 29.5%; 16–30 years, 13.6%; and > 31 years, 10%. Final spontaneous clearance rate of H. pylori after partial gastrectomy was 43%. For these backgrounds, the possibility of the spontaneous clearance of H. pylori (duration from 13C-UBT to endoscopic biopsies: 6 month to 1 year) would be caused the mismatch results.

For another important factor, regarding the role of bile reflux on H. pylori colonization in the remnant stomach after partial gastrectomy, many conflict results were published. According to a study by Onoda et al.,44 the prevalence of H. pylori infection was lower in Billroth-II reconstruction patients with severe bile reflux and subsequent stomal gastritis, suggesting a spontaneous eradication of H. pylori by the reflux of bile contents. Thus, subtotal gastrectomy allows bile reflux; they therefore cause more severe gastritis with decreased H. pylori infection. Although we could not evaluate the degree of bile reflux in the patients who had received subtotal gastrectomy, the reflux of bile acid might influence on H. pylori survival after 13C-UBT, and it might cause the mismatch result between the 13C-UBT and endoscopic biopsy methods.

However, a previous study reported that Roux-en-Y reconstruction after distal gastrectomy produces smaller amounts of bile reflux and as a result had a lower rate of H. pylori infection.45 Nakagawara et al.46 reported that bile refluxate facilitated the survival of H. pylori, speculating that H. pylori was perhaps inhibited by other bacteria in the gut. Pylorus-preserving gastrectomy for gastric cancer also resulted in significantly lower H. pylori prevalence after surgery.43 Thus, the precise mechanism between bile acid reflux and H. pylori survival would be required to further investigation.

Considering the accuracy of the 13C-UBT, Sheu et al.19 reported that the sensitivity and specificity of the UBT in the gastrectomy group were lower than those in the normal group. Applying a cut-off value as 2.5‰, the sensitivity and specificity were only 82.2% and 87.8%, respectively. They explained that such a poor UBT diagnostic efficacy in the gastrectomized patients can be attributed in part to the lower bacterial loads, inadequate coating of the stomach by urea, and disuse of test meal. Kubota et al.21 established a standardized protocol and cut-off value for the 13C-UBT in gastrectomized patients. Using receiver operating characteristic analysis, they selected 40 minutes and a cut-off of 2.0‰. As the delta over baseline (DOB) 13CO2 of H. pylori non-infected patients mainly reflect urease activity in the mouth, surgery should have no effect and the cut-off should be set appreciably higher.47 Graham et al.48 originally reported that a test meal was required before urea ingestion to similarly extend the period of contact between urea and H. pylori. Most studies evaluating the need for citric acid in UBT showed higher delta values with citric acid when compared with other test meals or no test meals.49–51 Citric acid is expected to increase delta values in infected patients and not change delta values in uninfected ones. Adding citric acid may, therefore, well increase the discriminative capacity of the test.49 In previous our reports,52 we also reported a high false positive results of the 13C-UBT after eradication of H. pylori in the normal patients. In line with previous reports, no application of the citric acid test meal for the test would be cause such a high rate of false positive results in this study.

In the present study, we also sought to identify clinical factors that caused false positive 13C-UBT results after H. pylori eradication in the remnant stomach. A previous history of multiple H. pylori eradication therapies was found to be correlated with false positive 13C-UBT results (odds ratio = 3.248, 95% confidence interval = 1.088–9.695, P = 0.035). Total 19 patients underwent second- or third-line eradication therapy for H. pylori infection in the 13C-UBT range over 2.5‰ in mismatched group. Of 2 patients who underwent third line eradication the 13C-UBT value were 6‰ and 22.4‰. In these cases, DOBs of the 13CO2 were decreased compared to first eradication therapy for H. pylori, but the change DOB 13CO2 range was variable. This consequence might imply that the diagnostic accuracy of the 13C-UBT is imprecise in positive value, thus additional second or third eradication would be needed according the UBT results. This is important point that if the clinicians should perform additional eradication of H. pylori infection more cautiously when the results fall into the positive results of the 13C-UBT after H. pylori eradication. Considering the altered stomach environment and the diagnostic reliability of the 13C-UBT, it would be better to postpone additional eradication and to perform endoscopic biopsy methods to detect for H. pylori infection in the remnant stomach.

However, this present study also has some limitations. First, we could not compare 13C-UBT and endoscopic biopsy results at the same time. According to the Korean Health Insurance service, these two tests should not be performed simultaneously after H. pylori eradication. Therefore, in this study, we evaluated H. pylori status endoscopically at least 6 months after the 13C-UBT. Thus, the high false positive results of the 13C-UBT might be influenced by the altered stomach environment, such as, increasing pH, bile reflux acid, and unused of test meal. Second, this study is intrinsically limited by its retrospective design. In particular, 42.3% of patients who performed 13C-UBT after receiving subtotal gastrectomy were enrolled, which introduce the possibility of sampling bias. Especially, the cases of the positive 13C-UBT result were lesser than half of negative results of the 13C-UBT. Third, the results of our study could be applied to post-H. pylori eradication status rather than initial diagnosis of H. pylori infection. Under the Korean National Medical Insurance system, it is difficult to obtain 13C-UBT values before H. pylori eradication. Furthermore, a well-designed, randomized, controlled study is needed to confirm the diagnostic validity of the 13C-UBT before H. pylori eradication in partial gastrectomized patients.

Summarizing, this study shows that there were too many mismatched results between the 13C-UBT and endoscopic biopsy methods after eradication of H. pylori infection in the patients who received partial gastrectomy for gastric cancer. Especially, in the range 2.5‰ to 10‰ of the 13C-UBT after H. pylori eradication in the gastrecomized patients, the clinicians should be consider the possibility of false positive result of 13C-UBT, and additional endoscopic surveillance with biopsy-based methods would be helpful to avoid unnecessary additional treatment for H. pylori infection.