|
 |
|
| |
|
|
|
| A new, practical, low-dose 14C-urea
breath test for the diagnosis of Helicobacter pylori infection:
clinical validation and comparison with the standard method |
Emel Özt?rk1, Zeki Yesilova2,
Seyfettin IIgan1, Nuri Arslan1, Ahmet Erdil2, B?lent Celasun3,
Mehmet Özg?ven1, Kemal Dagalp2, Önder Ovali4, Hikmet
Bayhan5
1Department of Nuclear Medicine, G?lhane Military Medical Academy
and Medical School, Etlik, Ankara, Turkey
2Department of Gastroenterology, G?lhane Military Medical Academy
and Medical School, Etlik, Ankara, Turkey
3Department of Pathology, G?lhane Military Medical Academy and
Medical School, Etlik, Ankara, Turkey
4Department of Internal Medicine, G?lhane Military Medical Academy
and Medical School, Etlik, Ankara, Turkey
5Department of Nuclear Medicine, Acibadem Hospital, Istanbul,
Turkey
Received: 1 May 2003 / Accepted: 10 May 2003 / Published online:
5 September 2003
©Springer – Verlag 2003 |
| |
| Abstract |
The carbon-14 urea breath test (UBT)
is a reliable and non-invasive technique for the diagnosis of
Helicobacter pylori (HP) infection. In this study we evaluated
the diagnostic performance of a new, practical and low-dose
14C-UBT system for the diagnosis of HP and compared
the results with those obtained using the standard method. Seventy-five
patients (56 female, 19 male) with dyspepsia underwent 14C-UBT
and endoscopy with antral biopsies for histological analysis.
The rapid urease test (CLO test) was applied to 50 of these
patients. After a 6-h fasting period, a 37-kBq 14C-urea
capsule was swallowed for UBT. Breath samples were collected
and counted using two different methods, the Heliprobe method
and the standard method. In the Heliprobe method, patients exhaled
into a special dry cartridge system (Heliprobe BreathCard) at
10 min. The activities of the cartridges were counted using
a designated small GM counter system (Heliprobe analyzer). Results
were expressed both as counts per minute (HCPM) and as grade
(0, not infected; 1, equivocal; 2, infected) according to the
counts. In the standard method, breath samples were collected
by trapping in a liquid CO2, absorber. Radioactivity
was counted as disintegrations per minute (SDPM) using a liquid
scintillation counter after addition of a liquid scintillation
cocktail.
Part of this work was presented at the 15th Congress of the
European Association of Nuclear Medicine, held on 31 August
to 4 September 2002 in Vienna, Austria.
Emel Özt?rk (?)
Department of Nuclear Medicine,
G?lhane Military Medical Academy and Medical School,
06018 Etlik, Ankara, Turkey
e-mail: docemelozturk@hotmail.com
Tel: +90-312-3044806, Fax: +90-312-3044800
|
Histological examination was used as
a gold standard. Two patients were excluded from the study because
of inadequate biopsy sampling. Forty-eight patients (65%) were
found to be HP positive on histology. The Heliprobe method correctly
classified 48 of 48 HP-positive patients and 19 of 25 HP-negative
patients (sensitivity 100%, specificity 76%, PPV 88%, NPV 100%,
accuracy 91%). The standard method correctly classified 48 of
48 HP-positive and 20 of 25 HP-negative patients (sensitivity
100%, specificity 80%, PPV 90%, NPV 100%, accuracy 93%). On
the other hand, the CLO test identified 26 of 32 HP-positive
and 12 of 16 HP-negative patients (sensitivity 81%, specificity
75%, PPV 86%, NPV 66%, accuracy 79%). With the Heliprobe method,
all of the positive results were grade 2, and all of the negative
results were grade 0. No patients were defined as having grade
1 results. Counts allowed clear discrimination of HP-positive
and –negative patients with both methods, the difference
being statistically significant in each case (P<0.001). A
significant correlation was found between HCPM and SDPM (r 0.863,
P<0.001). According to the ROC analysis, the area under the
curve was nearly the same with HCPM (AUC, 0.888; 95% CI, 0.785-0.992)
and SDPM (AUC, 0.898; 95% CI, 0.802-0.994). In conclusion, the
new 14C-UBT system is a highly accurate method for the diagnosis
of HP infection. It is rapid and practical, and therefore suitable
for clinical and office practice.
Keywords: Helicobacter pylori – Carbon -14 urea breath
test- Peptic ulcer disease.
Eur J Nucl Med Mol Imaging (2003) 30:1457-1462
DOI 10.1007/s00259-003-1244-8 |
| |
|
top |
| Introduction |
Helicobacter pylori (HP) is a spiral,
gram-negative bacterium that has been found to be associated
with gastritis, peptic ulcer disease, gastric adenocarcinoma
and MALT lymphoma [1,2,3]. There is consequently increasing
demand for treatment and a great need for simple and accurate
methods for the diagnosis of HP infection.
Invasive diagnostic methods require mucosal biopsy during endoscopy,
with the specimens being subjected to culture, rapid urease
test, polymerase chain reaction or histological analysis. Non-invasive
methods include antibody detection (serology), stool antigen
test and urea breath test (UBT). While serology (ELISA) is simple
and easy to perform, it is not a reliable test for the diagnosis
of HP infection in elderly people or for determination of eradication
of HP, since it remains positive for a long period despite adequate
treatment [4, 5].
The production of high amounts of urease by HP has been used
in the development of UBTs. An oral dose of urea is rapidly
broken down by HP in the gastric mucosa to ammonia and carbon
dioxide. Labeled carbon dioxide derived from labeled urea can
be detected in the breath as a marker of infection. UBTs with
either carbon-13 or carbon-14 urea are non-invasive methods,
sample the whole stomach and reflect the actual status of infection.
Both tests are highly accurate, with reported sensitivities
of 97-100% and specificities of 95-100% for both diagnosis and
proof of eradication of HP infection after therapy [1, 2, 6,
7, 8, 9]. While the two isotopes seem to offer similar diagnostic
accuracy, 13C-UBT has the inconvenience of requiring (a) more
complex and expensive equipment on site or else analysis off-site
by an external laboratory and (b) administration of a test meal
and cold urea to the patient. These are not necessary with 14C,
and the test is thus simpler, faster and cheaper.
The routine test protocol of 14C-UBT requires ingestion
of 14C-urea, collection of breath samples at frequent
intervals using a liquid CO2, trapping medium, addition of a
liquid scintillation cocktail and counting with a ß–scintillation
counter. Several different methodological approaches have been
suggested to simplify the UBT. Recently a new, practical dry
breath collection cartridge (Heliprobe BreathCard) and counting
system (Heliprobe analyzer) have been developed for this purpose.
In this prospective study we evaluated the diagnostic performance
and accuracy of this new 14C-UBT system and compared
the results with those of the rapid urease test (CLO test) and
the standard (liquid CO2 absorber and liquid scintillation
counting) method. |
| |
| Materials and Methods |
Patients. Seventy-five patients (56
female, 19 male; mean age 41±14 years) with dyspepsia
were included in the study. Informed consent was obtained from
each patient. All patients underwent upper gastrointestinal
endoscopy as well as 14C-UBT within 1 week. CLO test
was applied to 50 of these patients.
Rapid urease test (CLO) test and histological examination. During
upper gastrointestinal endoscopy, three biopsy specimens were
taken from antral mucosa, one for CLO test and two for histological
analysis. A home made kit was used for the CLO test. One biopsy
specimen was placed in a test tube and the colour reaction was
read after 6 h. Histology samples were fixed in formalin, embedded
in paraffin, sectioned in routine fashion and stained with Giemsa.
14C-urea breath test. Antacids were stopped
at least 24 h before the test, sucralfate and H2, receptor antagonists
were discontinued for 1 week before the test, and proton pump
inhibitors, bismuth compounds and antibiotics were stopped for
1 month beforehand. After overnight fasting, patients swallowed
37 kBq (1 µCi) of an encapsuled form of 14C-urea/citric
acid composition (Helicap, Noster System AB Stockholm, Sweden)
with 25 ml water. Breath samples were collected and counted
using two different methods. |
| 1. |
Heliprobe method:
Breath samples of patients were collected with a special
dry cartridge system (Heliprobe BreathCard. Noster System
AB Stockholm, Sweden) at 10 min. Patients exhaled gently
into the cartridge mouthpiece until the indicator membrane
changed colour from orange to yellow. The breath-card
was inserted into a special small desktop Geiger-M?ller
counter (Heliprobe-analyzer, Noster System AB Stockholm,
Sweden) and activity counted for 250 s. Results were expressed
both as counts per minute (HCPM) and as grade (0: not
infected, CPM <25:1:equivocal, CPM 25-50; 2: infected,
CPM >50), as suggested by the producer according to
the counts obtained from the cartridges. |
| 2. |
Standard method: After completion
of the breath sample collection in method 1, patients
were asked to blow through a drinking straw into a 20-ml
glass scintillation vial containing 0.1 ml CO2,
absorber solution (Carba Sorb E, Packard) as well as a
trace of the pH indicator thymolphthalein. Sampling was
considered complete when the colour of the solution changed
from blue to colourless. After addition of 10 ml of liquid
scintillation cocktail (Pico-flour 40, Packard), radioactivity
was counted for 10 min in a liquid scintillation counter
(tri-carb 2500 TR, Packard). Counts were corrected to
disintegrations per minute (SDPM) using an external standard
method. The cut-off value was chosen as 100 DPM for the
standard method in accordance with the results of our
previous biopsy-controlled study. |
|
|
top |
Data analysis and statistics. Both 14C-UBT
methods and the CLO test were validated against histological
examination, and their sensitivity, specificity, positive/negative
predictive values (PPV, NPV) and accuracy were determined.
The HCPM and SDPM counts of HP-positive and –negative
patients were compared using the Mann-Whitney U test. Spearman’s
correlation and ROC curve analysis were performed for comparison
of HCPM and SDPM counts.
All statistical analyses, except for the diagnostic performance
values, were performed with Systat (ver.10) statistical package
(SPSS, Chicago, IL)
|
| Table 1. Comparative results of
histology, the CLO test, the Heliprobe method and the standard
method |
| Histology |
CLO test |
Heliprobe UBT |
Standard UBT |
| |
(+) |
(-) |
(+) |
(-) |
(+) |
(-) |
| HP (+) |
26 |
6 |
48 |
- |
48 |
- |
| HP (-) |
4 |
12 |
6 |
19 |
5 |
20 |
|
| |
| Table 2. Diagnostic performance
of the tests |
| |
CLO test |
Heliprobe UBT |
Standard UBT |
| Sensitivity (%) |
81 |
100 |
100 |
| Specificity (%) |
75 |
76 |
80 |
| Positive Predictive value
(%) |
86 |
88 |
90 |
| Negative Predictive value
(%) |
66 |
100 |
100 |
| Accuracy (%) |
79 |
91 |
93 |
|
| |
| Table 3. The HCPM and SDPM values
of Helicobacter pylori positive and –negative patients |
| |
HP (+) |
HP (-) |
| HCPM |
269 (300, 69-770)3 |
10 (72, 0-617) |
| SDPM |
745 (1,104,220-3, 747) |
34 (216, 4-1, 422) |
|
| 3 Median (mean, minimum–maximum) |
| |
|
| |
| Results |
| Two patients were excluded from the
study because of inadequate biopsy sampling. On histology, 48
patients (65%) were found to be HP Positive, and 25 HP negative. |
| |
| CLO test |
| Among the 48 patients evaluated with
CLO test, 26 of 32 HP-positive (sensitivity81% and PPV 86%)
and 12 of 16 HP–negative patients (specificity 75% and
NPV 66%) were correctly identified. Accuracy was 79% (Tables
1 and 2) |
| |
|
| |
| Heliprobe and standard methods |
Table 3 shows the spread of the counts
with both methods. Counts allowed clear discrimination of HP-positive
and –negative patients with both methods, the difference
being statistically significant in each case (P<0.001).
The Heliprobe method correctly classified 48 of 48 HP-positive
patients and 19 of 25 HP-negative patients (sensitivity 100%,
specificity 76%). The standard method correctly classified 48
of 48 HP-positive patients and 20 of 25 HP-negative patients
(sensitivity 100%, specificity 80%) (Tables 1 and 2).
According to the grading with the Heliprobe method, all of the
positive results were grade 2 (infected) and all of the negative
results were grade 0 (not infected). No patients were defined
as having grade 1 (equivocal) results. |
| |
 |
| Fig. 1. Comparison of SDPM and
HCPM |
| |
|
| |
| Heliprobe method versus standard
method |
A significant correlation was found
between HCPM and SDPM counts (r 0.863, P<0.001, Fig.1). According
to the ROC analysis, the area under the curve was nearly the
same with HCPM (AUC, 0.888; 95% CI, 0.785-0.992) and SDPM (AUC,
0.898; 95% CI, 0.802-0.994) (Fig.2).
Three HP-negative patients showed discordant results with the
standard and Heliprobe methods. Two of them had positive test
results only with the Heliprobe method (pt.6: HCPM 105, SDPM
52; pt.8: HCPM 52, SDPM 22),whereas one of them was positive
only with the standard method (pt.4: SDPM 330, HCPM 21). |
| |
|
|
top |
| |
 |
| Fig. 2. ROC curve analysis of
HCPM and SDPM |
| |
| |
|
|
|
top |
|
