Original Paper

Med Princ Pract 2009;18:57–61 Received: February 19, 2008

Revised: July 6, 2008
DOI: 10.1159/000163048

Comparison of Noninvasive Diagnostic

Tests for Helicobacter pylori Infection
Nan-Jing Peng a, c Kwok-Hung Lai b, c Gin-Ho Lo b, c Ping-I Hsu b, c
a
Department of Nuclear Medicine and b Division of Gastroenterology, Department of Internal Medicine,
Kaohsiung Veterans General Hospital, Kaohsiung; cNational Yang-Ming University, Taipei, Taiwan, ROC

Key Words According to our study, capsule UBT was highly accurate
Capsule 13C-urea breath test ⴢ Helicobacter pylori ⴢ Serology compared with other noninvasive tests including conven-
tional UBT and serology. It could become a good alternative
to endoscopy for the diagnosis of H. pylori infection.
Abstract Copyright © 2008 S. Karger AG, Basel
Objectives: Since the 13C-urea breath test (UBT) has become
a highly reliable method for the noninvasive diagnosis of He-
licobacter pylori infection, this study was performed in order Introduction
to compare the sensitivity, specificity and accuracy among
noninvasive tests including capsule UBT, conventional UBT This study was performed in order to compare the
and serology in the diagnosis of H. pylori infection. Patients sensitivity, specificity and accuracy among noninvasive
and Methods: One hundred patients received capsule UBT, tests including capsule 13C-urea breath test (UBT), con-
conventional UBT and gave blood samples for the diagnosis ventional UBT and serology in the diagnosis of Helico-
of H. pylori infection. Upper gastrointestinal endoscopy was bacter pylori infection. H. pylori infection can be diag-
performed in all patients. H. pylori infection was defined as nosed by invasive and noninvasive techniques. Invasive
the presence of a positive culture or positive results of both methods require endoscopy- and biopsy-based tests, in-
histology and rapid urease test (CLO test). McNemar’s test cluding microbiological culture, histology, rapid urease
was used to determine the significance of differences among test (CLO test) and polymerase chain reaction. Noninva-
capsule UBT, conventional UBT and serology. Differences sive tests include serology, stool antigen test and breath
were considered significant at p ! 0.05. Results: According test. The choice of a diagnostic test should depend on the
to the predefined criteria, the sensitivity, specificity, positive clinical circumstances, sensitivity and specificity of the
predictive value and negative predictive value of capsule tests, and the cost-effectiveness of the testing strategy.
UBT, conventional UBT and serology was 100, 95.7, 96.4 and According to our prior articles [1, 2], the sensitivity of
100%; 100, 85.1, 88.3 and 100%, and 90.6, 85.1, 82.7 and culture, histology and CLO test was 77.8–94.4, 88.9–90.9
88.9%, respectively. The accuracy of capsule UBT was higher and 82.1–94.4%, respectively, and the specificity was 100,
than that of conventional UBT and serology (98 vs. 93 and 90.6–100 and 95.5–96.9%, respectively. The overall accu-
88%, respectively). Capsule UBT had a similar ability for the racy of serological assays average 78% [3]. The sensitivity
detection of H. pylori infection compared with conventional and specificity of UBT have been shown to range from 90
UBT and serology (McNemar’s test, p 1 0.05). Conclusions: to 100%, compared with biopsy-based tests for H. pylori

© 2008 S. Karger AG, Basel Ping-I Hsu

1011–7571/09/0181–0057$26.00/0 Division of Gastroenterology, Department of Internal Medicine
Fax +41 61 306 12 34 Kaohsiung Veterans General Hospital
E-Mail karger@karger.ch Accessible online at: 386, Ta-Chung 1st Road, Kaohsiung 813, Taiwan (ROC)
www.karger.com www.karger.com/mpp Tel. +886 7 342 2121 2075, Fax +886 7 346 8234, E-Mail williamhsup@yahoo.com.tw
infection [4]. Because of the excellent diagnostic accura- tive oxidase, catalase and urease reaction. Histology-positive pa-
cy, this test has been proposed for replacing endoscopy to tients were those with curved organisms seen in hematoxylin and
eosin-stained sections under the microscope. Although there is
diagnose H. pylori infection [5–8]. It is also an ideal meth- no real gold standard for the diagnosis of H. pylori, cultures
od in monitoring treatment success following anti-H. py- should be 100% specific if the procedures are performed properly.
lori therapy, because it can avoid the false-negative result Use of culture alone as the gold standard may yield false-negative
from biopsy-based tests due to focal colonization of bac- results due to the inherent difficulties of culture. Hence, in the
teria and the false-positive result from serology due to the present study we defined patients with H. pylori infection as those
with positive culture or positive results from histology and CLO
presence of anti-H. pylori IgG antibody [9]. test.
UBT is based on the high urease activity of H. pylori
in vivo. The orally administered 13C-labeled urea is read- Conventional UBT and Capsule UBT
ily hydrolyzed by the urease of the organism to ammonia Conventional UBT and capsule UBT were performed 2 days
and 13CO2. The latter can be detected through collection apart within 7 days after endoscopy before any antibiotic treat-
ment was given. No test meal was used in these tests. The patients
of expired samples. We recently reported a trial of a cap- were asked to fast at least 6 h. The 13C-urea, 100 mg of 99 atom %
sule UBT for the detection of H. pylori infection [2]. This of 13C-labeled urea, was supplied with a commercial package from
diagnostic method may avoid contamination of urea Japan (UBiT). On conventional UBT, the 13C-urea solution was
from oral urease. Additionally, it can rapidly discrimi- prepared by dissolving 100 mg 13C-urea in 50 ml sterile water in
nate between H. pylori-positive and H. pylori-negative a drinking vessel. For capsule UBT, a 100-mg capsule of 13C-urea
was packed in a gelatin capsule. Patients drank the 100-mg 13C-
patients with 100% sensitivity and specificity. Its diag- urea solution. Immediately after 13C-urea consumption, patients
nostic accuracy was significantly higher than that of cul- underwent a mouth washing by gargling to avoid oral 13C-urea
ture, histology, and CLO test. activity. Breath samples were collected before and 15 min after
consumption of 13C-urea. All breath samples were collected
breathing out to a 200-ml gas storage bag for infrared spectrom-
eter. The excess 13CO2/12CO2 ratio samples were analyzed by an
Patients and Methods infrared spectrometer at our department (UBiT-IR300, Photal
Otsuka Electronics Co., Japan). According to our previous studies
Patients [2, 8], the cutoff values were set at 4.8‰ for conventional UBT and
One hundred patients who presented for routine upper gastro- 2‰ for capsule UBT to judge the agreement of H. pylori-positive
intestinal endoscopy and were willing to cooperate with this and H. pylori-negative. The conventional UBT used in this study
study were included in the present study. Fifty-six were male and is identical to our previous study [8] except for the abolishment of
44 female, with a mean age of 55 years (range, 18–83 years). Cri- the test meal because the test meal did not affect UBT results at
teria for exclusion included (1) use of proton pump inhibitors 15 min, but increased values at 30 min and thereafter [10].
within 1 month before endoscopy; (2) antibiotic ingestion within Values were expressed as an excess ␦13CO2 ‰ excretion. The
1 month before endoscopy; (3) serious medical illness, and (4) ␦ CO2 is the ratio of 13C to 12C in the sample compared to the Pee
13

previous history of anti-H. pylori therapy. The study was approved Dee Belemnite standard. The equation is given as: ␦13CO2 = (Rsamp
by the institutional review board and the hospital’s ethics com- – Rstd)/Rstd ! 1,000. Rsamp and Rstd represent the ratios of 13C to
mittee. All participants gave written informed consent. 12
C in sample and standard, respectively. Excess ␦13CO2 is the
value of ␦13CO2 detected at 15 min minus that at baseline.
Upper Gastrointestinal Endoscopy, Culture, Histology and
CLO Test Serology Test
During endoscopy, four gastric biopsy specimens were taken Blood samples for serologic evaluation were taken in the
from the lesser curvature 2 cm from the pylorus. Two specimens morning of the first UBT test. A serological assay for IgG anti-
were fixed immediately in 10% neutral buffered formalin for his- bodies against H. pylori was performed by a commercial test kit
tological examination, one for culture and the other for urease test (ASSURETM H. pylori rapid test kit, Genelabs Diagnostics, Caven-
(CLO test, Delta-West, Bently, Australia). Samples were sent to dish Singapore Science Park, Singapore). The test is an indirect
different laboratories that were blind to the results of other tests. solid-phase immunochromatographic assay where antibodies in
The CLO test was monitored for color change up to 24 h after the the test sample form antibody-antigen complexes with immobi-
addition of the gastric tissue. The gel was not warmed above am- lized H. pylori antigens on the membrane as the test samples mi-
bient temperature at any time during the incubation period. The grate upwards from the sample well. The bound antibody-antigen
specimens for microbiological examination were transferred with complexes are subsequently detected by anti-human IgG conju-
brain-heart infusion in ice and inoculated onto the CDC anaero- gated to colloidal gold. In brief, 25 ␮l of a whole blood sample was
bic blood agar (Becton Dickinson Microbiology System, Cock- added to a sample well followed by 1 drop of chase buffer to the
eysville, Md., USA). The agar was incubated at 35 ° C for 7 days in well. When the sample front reached the pink control line, three
a microaerophilic gas mixture composed of 5% O2, 10% CO2, and drops of chase buffer were added to the buffer well. If colored
85% N2. Culture-positive patients were those with bacterial colo- bands appeared at (1) all the control line, CIM line and test line,
nies grown in culture within 7 days. The organisms were identi- or (2) both the control line and test line, the test was regarded as
fied as H. pylori by Gram staining, colony morphology and posi- positive. If only the control line was visible, the test was regarded

58 Med Princ Pract 2009;18:57–61 Peng /Lai /Lo /Hsu

Table 1. Comparison of the sensitivity,
specificity, PPV, NPV and accuracy of Serology Conventional UBT Capsule UBT
capsule UBT, conventional UBT and
serology in the diagnosis of H. pylori Sensitivity (CI) 90.6 (82.7–98.5) 100 (100–100) 100 (100–100)
infection expressed as percentages Specificity (CI) 85.1 (74.9–95.3) 85.1 (74.8–95.2) 95.7 (89.9–100)
PPV (CI) 87.2 (78.5–96.1) 88.3 (80.2–96.4) 96.4 (91.5–100)
NPV (CI) 88.9 (79.7–98.1) 100 (100–100) 100 (100–100)
Accuracy (CI) 88 (81.6–94.4) 93 (88–98) 98 (95.3–100)

Table 2. Matched reading of capsule UBT versus conventional ventional UBT (96.4 vs. 88.3 and 87.2%, respectively). The
UBT and serology NPV of the capsule UBT was higher than that of serology
(100 vs. 88.9%), and equal to that of conventional UBT
Capsule UBT Conventional UBT Serology
(100 vs. 100%). The capsule UBT had a higher accuracy
positive negative positive negative compared with conventional UBT and serology (98 vs. 93
and 88%).
Positive 55 5 50 5
Negative 0 40 5 40
Table 2 shows the matched reading of capsule UBT
versus conventional UBT and serology. Based on the two-
McNemar’s test p = 0.063 p=1 tailed McNemar’s test, capsule UBT detected the same
number of positive and negative H. pylori as serology
(p = 1), and a slightly different number compared with
conventional UBT (p = 0.063). Therefore, capsule UBT
as negative. If the control line was absent or both control line and
CIM line were present but the test line absent, the test was regard- had similar ability for the detection of H. pylori infection
ed as invalid. compared with conventional UBT and serology (McNe-
mar’s test, p 1 0.05).
Statistical Analysis
Sensitivity, specificity and accuracy of capsule UBT, conven-
tional UBT and serology were calculated according to the pre-
defined gold standard. Proportions of positive and negative re- Discussion
sults of capsule UBT, conventional UBT and serology for the di-
agnosis of H. pylori infection were compared using the two-tailed Conventional UBT with administration of liquid solu-
McNemar’s test for matched pairs [11]. A p value less than 0.05 tion has been regarded as an accurate method for the
was considered as statistically significant. Ninety-five percent noninvasive diagnosis of H. pylori infection. Its sensitiv-
confidence intervals (CIs) were also calculated. Narrow and high
CIs are desirable. Wide and low CIs indicate poor agreement [12]. ity and specificity have been shown to range from 90 to
All calculations were performed using SPSS version 12.0 (SPSS 100%, compared with biopsy-based tests [4]. Most inves-
Inc., Chicago, Ill., USA). tigators collected the breath samples at least 10 min after
consumption because urease from oral flora may cause
false-positive results if breath samples are taken too soon
Results [4]. According to our previous studies and this study, the
false-positive rates of conventional UBT still ranged from
In the total of 100 patients, 53 were H. pylori-positive 4.3 to 14.9% [8, 9]. The decrease in specificity of conven-
and 47 H. pylori-negative according to the predefined cri- tional UBT may be related to a number of factors. Urease
teria. Sensitivity, specificity, positive predictive value from the oral bacterial flora is an important false-positive
(PPV), negative predictive value (NPV) and accuracy of factor in the breath test for H. pylori. Cleansing the pa-
the capsule UBT, conventional UBT and serology are tients’ mouths and delayed sample collection have com-
shown in table 1. The sensitivity of the capsule UBT was monly been used to avoid this factor. Tests performed at
higher than that of serology (100 vs. 90.6%), and equal to 15 min were less specific than those performed at 30 min,
that of conventional UBT (100 vs. 100%). The specificity probably because of the interference of oral urease-pro-
of the capsule UBT was higher than that of conventional ducing organisms. Another factor is the cutoff value used.
UBT and serology (95.7 vs. 85.1 and 85.1%, respectively). Based on our previous studies, we chose 4.8‰ as the cut-
The PPV of the capsule UBT was higher than that of con- off value and achieved 100% sensitivity and 85.1% speci-

Noninvasive Tests for H. pylori Infection Med Princ Pract 2009;18:57–61 59

ficity on conventional UBT. Increased cutoff values might sibility of false-negative results from biopsy-based tests.
achieve higher specificity, but lower sensitivity. That is also the reason why conventional UBT has been
We have previously demonstrated that the timing of proposed for replacing endoscopy to diagnose H. pylori
sample collection can be shortened to 2 min with 100% infection [5–8]. In the present study, capsule UBT dem-
sensitivity and specificity through endoscopic adminis- onstrated higher accuracy than conventional UBT. In ad-
tration [1, 13]. However, endoscopy is invasive, expensive, dition, the 13C-urea solution was prepared by dissolving
causes patient discomfort, and introduces the risk of en- 100 mg 13C-urea in 50 ml sterile water in a drinking ves-
doscopy cross-infection. Patient-to-patient transmission sel for conventional UBT. For capsule UBT, a 100-mg cap-
of H. pylori infection by gastroduodenoscopy may result sule of 13C-urea was packed in a gelatin capsule. Applica-
from the biopsy procedure or the insertion of an endo- tion of the capsule UBT shortened the preparation time
scope that is not properly cleaned [14, 15]. Because the and avoided contamination during preparation. We
biopsy forceps are usually reused after sterilization in therefore suggest the capsule UBT as a good alternative
Taiwan and in many developing countries, gastric biopsy to endoscopy and conventional UBT as the gold standard
may increase the risk of cross-infection. Application of for the diagnosis of H. pylori infection.
encapsulated urea has the same effect as endoscopic UBT The accuracy of our serological test was 88%. In an
to prevent the release of urea before reaching the stomach independent study by Rahman et al. [17], the accuracy of
and avoids the contamination of oral urease. We have re- the test was also 88%. In our study, the accuracy of cap-
cently reported a trial of a capsule UBT for the detection sule UBT was higher than that of serology (98 vs. 88%).
of H. pylori infection [2]. This diagnostic method avoids The values of capsule UBT are able to reflect the bacte-
contamination of urea from oral urease. Additionally, it rial load in the stomach, whereas those of serology are
can rapidly discriminate between H. pylori-positive and not. Several authors have proposed a correlation between
H. pylori-negative patients. Its diagnostic accuracy was UBT values and H. pylori bacterial load [18–20]. Suto et
significantly higher than that of culture, histology, and al. [20] have found positive correlations between endo-
CLO test [2]. In the present study we compared the cap- scopic UBT values and H. pylori colonization and activ-
sule UBT with conventional UBT in the same patients ity score in the antrum and corpus, and negative correla-
and showed that the capsule UBT has a similar ability for tions between the endoscopic UBT values and the atro-
the detection of H. pylori infection compared with con- phy and intestinal metaplasia scores in the corpus.
ventional UBT (McNemar’s test, p = 0.063) and a higher Endoscopic and capsule UBT theoretically are the most
accuracy than conventional UBT (98 vs. 93%). accurate methods to quantitatively assess bacterial load
According to the predefined criteria, there were 2 in the stomach since they avoid the influence of urease
false-positive capsule UBTs. Conventional UBT and se- activity in the oral cavity. Furthermore, capsule UBT is
rology were also found to be positive in these 2 subjects. also an ideal method for monitoring treatment success
We doubted the diagnostic accuracy of predefined crite- following anti-H. pylori therapy, where serology is inval-
ria because biopsy-based tests may lead to false-negative id due to the presence of anti-H. pylori IgG antibody [9].
results from suboptimal sampling due to focal coloniza- The presence of anti-H. pylori IgG antibody implies prior
tion of bacteria. Based on our prior article, the sensitivity exposure to these organisms, but does not imply the pres-
of culture, histology and CLO test were generally around ence of a current infection. Therefore, capsule UBT is su-
77.8–94.4, 88.9–90.9 and 82.1–94.4%, respectively [1, 2], perior to serology for noninvasive diagnosis of H. pylori
but 66, 43 and 37% in patients with bleeding peptic ulcer, infection.
respectively [16]. However, the sensitivity of serology and In this study, we demonstrated that capsule UBT was
UBT was both 94% in patients with bleeding peptic ulcers an extremely accurate method for the diagnosis of H. py-
[16]. Since patients with peptic ulcer bleeding were not lori infection. Because urease is not present in a healthy
excluded in the study, the false-positive UBTs might be stomach, capsule UBT rarely causes a false-positive re-
due to the low sensitivity of the biopsy-based tests in pa- sult. UBT effectively integrates the enzyme activity of
tients with bleeding peptic ulcers with which the UBT has bacteria over the entire surface of the stomach, and avoids
been compared. Therefore, the low sensitivity of biopsy- the possibility of false-negative results. Capsule UBT pro-
based tests in patients with bleeding peptic ulcers may vides an excellent method both for the initial diagnosis
increase the proportion of false-positive results in UBTs. of H. pylori infection and for the confirmation of its erad-
UBT effectively integrates the enzyme activity of bacteria ication after treatment. Nowadays, the strategy of ‘test
over the entire surface of the stomach, and avoids the pos- and treat’ (diagnose the H. pylori infection and treat it)

60 Med Princ Pract 2009;18:57–61 Peng /Lai /Lo /Hsu

was recommended for patients with dyspeptic symptoms. compared with conventional UBT and serology in the
However, failure of ‘test and treat’, 1 45 years of age or the present study. It could be a good alternative to endoscopy
presence of alarm features still deserve prompt endos- for the diagnosis of H. pylori infection.
copy. Although the diagnostic accuracy of capsule UBT
is very high, it still cannot replace endoscopy. It can be
used along with invasive tests to improve diagnostic ac- Acknowledgment
curacy and in all conditions not requiring endoscopy.
This work was supported by grant VGHKS94-93 from Kaoh-
Our study showed that capsule UBT is a simple, nonin-
siung Veterans General Hospital, Taiwan, Republic of China.
vasive, and extremely accurate method for the diagnosis
of H. pylori infection. The new test had a high accuracy

References

1 Peng NJ, Lai KH, Liu RS, Lee SC, Tsay DG, 8 Peng NJ, Hsu PI, Lee SC, Tseng HH, Huang 15 Miyaji H, Kohli Y, Azuma T, Ito S, Hirai M,
Lo CC, Tseng HH, Huang WK, Lo GH, Hsu WK, Tsay DG, Ger LP, Lo GH, Lin CK, Tsai Ito Y, Kato T, Kuriyama M: Endoscopic
PI: Endoscopic 13C-urea breath test for the CC, Lai KH: A 15-minute 13C-urea breath cross-infection with Helicobacter pylori.
diagnosis of Helicobacter pylori infection. test for the diagnosis of Helicobacter pylori Lancet 1995;345:464.
Dig Liver Dis 2003;35:73–77. infection in patients with non-ulcer dyspep- 16 Lo CC, Lai KH, Peng NJ, Lo GH, Tseng HH,
2 Peng NJ, Lai KH, Liu RS, Lee SC, Tsay DG, sia. J Gastroenterol Hepatol 2000; 15: 284– Lin CK, Shie CB, Wu CM, Chen YS, Huang
Lo CC, Tseng HH, Huang WK, Lo GH, Hsu 289. WK, Chen A, Hsu PI: Polymerase chain reac-
PI: Capsule 13C-urea breath test for the diag- 9 Peng NJ, Hsu PI, Lee SC, Tseng HH, Huang tion: a sensitive method for detecting Helico-
nosis of Helicobacter pylori infection. World WK, Tsay DG, Lo GH, Lin CK, Cheng JS, Lai bacter pylori infection in bleeding peptic ul-
J Gastroenterol 2005; 11:1361–1364. KH: Follow-up of Helicobacter pylori status cers. World J Gastroenterol 2005; 11: 3909–
3 Stevens M, Livsey S, Swann R: Evaluation of by using 13C-urea breath test in nonulcer 3914.
Sixteen EIAs for the Detection of Antibodies dyspeptic patients after eradication therapy. 17 Rahman SHZ, Azam MG, Radman MA, Ar-
to Helicobacter pylori. London, Department Chin Med J (Taipei) 2001;64:337–342. fin MS, Alam MM, Bhuiyan TM, Ahmed N,
of Health, 1997, pp 1–46. 10 Atherton JC, Washington N, Blackshaw PE, Rahman M, Nahar S, Hassan MS: Non-inva-
4 Gisbert JP, Pajares JM: Review article: 13C- Greaves JL, Perkins AC, Hawkey CJ, Spiller sive diagnosis of H pylori infection: evalua-
urea breath test in the diagnosis of Helico- RC: Effect of a test meal on the intragastric tion of serological tests with and without
bacter pylori infection – a critical review. Al- distribution of urea in the 13C-urea breath current infection marker CIM. World J Gas-
iment Pharmacol Ther 2004;20:1001–1017. test for Helicobacter pylori. Gut 1995;36:337– troenterol 2008;28:1231–1236.
5 Slomiansk A, Shubert T, Cutler AF: [13C]urea 340. 18 Perri F, Clemente R, Pastore M, Quitadamo
breath test to confirm eradication of Helico- 11 Fleiss J: Statistical Methods for Rates and M, Festa V, Bisceglia M, Li Bergoli M, Lau-
bacter pylori. Am J Gastroenterol 1995; 90: Proportions, ed 2. New York, Wiley, 1981, pp riola G, Leandro G, Ghoos Y, Rutgeerts P,
224–226. 113–115. Andriulli A: The 13C-urea breath test as a
6 Epple HJ, Kirstein FW, Bojarski C, Frege J, 12 Grant JM: The fetal heart rate trace is nor- predictor of intragastric bacterial load and
Fromm M, Riecken EO, Schulzke JD: 13C- mal, isn’t it? Observer agreement of categor- severity of Helicobacter pylori gastritis.
urea breath test in Helicobacter pylori diag- ical assessments. Lancet 1991; 337:215–218. Scand J Clin Lab Invest 1998;58:19–28.
nosis and eradication: correlation to histol- 13 Peng NJ, Lai KH, Liu RS, Lee SC, Tsay DG, 19 Debongnie JC, Pauwels S, Raat A, de Meeus
ogy, origin of ‘false’ results, and influence of Lo CC, Tseng HH, Huang WK, Lo CC, Hsu Y, Haot J, Mainguer P: Quantification of He-
food intake. Scand J Gastroenterol 1997; 32: PI: Clinical significance of oral urease in di- licobacter pylori infection in gastritis and ul-
308–314. agnosis of Helicobacter pylori infection by cer disease using a simple and rapid carbon-
7 Katelaris PH, Jones DB: Testing for Helico- 13C-urea breath test. Dig Dis Sci 2001; 46: 14-urea breath test. J Nucl Med 1991; 32:
bacter pylori infection after antibiotic treat- 1772–1778. 1192–1198.
ment. Am J Gastroenterol 1997; 92: 1245– 14 Langenberg W, Rauws EAJ, Oudbier JH, Tyt- 20 Suto H, Azuma T, Ito S, Ito Y, Miyaji H,
1247. gat GNJ: Patient-to-patient transmission of Yamazaki Y, Kohli Y, Kuriyama M: Endo-
Campylobacter pylori infection by fiberoptic scopic 13C-urea breath test for quantification
gastroduodenoscopy and biopsy. J Infect Dis of Helicobacter pylori infection. J Gastroen-
1990;161:507–511. terol Hepatol 2000; 15:161–167.

Noninvasive Tests for H. pylori Infection Med Princ Pract 2009;18:57–61 61