voting slides. The statements were presented at the World Congress
of Pediatric Gastroenterology in Iguassu Falls, Brazil, on August
19, 2008 to the scientific community and feedback was requested.
The first-draft manuscript was prepared by the chair of the
European group, Sibylle Koletzko, in collaboration with Nicola
Jones of the North American group, and the 2 epidemiologists
Karen Goodman and Marion Rowland. Because of a change in the
NASPGHAN chair, the manuscript was on hold for 18 months. In
December 2009, an updated systematic literature search was per-
formed including articles published from September 2007 to 2009.
A total of 248 new publications were retrieved and reviewed for
new evidence, which may have influence on the recommendations,
the evidence, or the strength of recommendations compared
with the version presented in August 2008 at the World Congress.
The new literature was implemented in the final draft, which then
was circulated to all members of the consensus group and their input
was worked into the manuscript.
3. RESULTS
3.1. Statements and Comments
For the first round of voting, 43 statements were presented
and agreement was reached for 22 of them. Several statements
were omitted, some combined into 1, and others reworded after
discussion. There were 21 statements in the final round of voting,
and consensus was reached for all of them. The result of the final
voting is provided for every statement.
3.2. Who Should Be Tested?
Recommendation 1
The primary goal of clinical investigation
of gastrointestinal symptoms is to determine the underlying
cause of the symptoms and not solely the presence of H pylori
infection.
Agree: 100% (AR 92%, A 8%). Grade of evidence: not
applicable.
Recommendation 2
Diagnostic testing for H pylori infection is
not recommended in children with functional abdominal pain.
Agree: 92% (AR 54%, A 23%, AS 15%, DS 8%). Grade
of evidence: high.
Comment on Recommendations 1 and 2:
Abdominal complaints such as pain, nausea, or other
dyspeptic symptoms are nonspecific and can be caused by different
organic diseases within and outside the digestive tract. These
diseases may be missed or their diagnosis and treatment delayed,
if a noninvasive test for H pylori infection is positive and treatment
initiated. For example, Levine et al (17) performed endoscopy in
children with epigastric pain and symptoms suggestive of gastro-
esophageal reflux disease. After treatment, improvement of
epigastric pain correlated with improvement of reflux disease,
but was not related to H pylori eradication. Abdominal complaints
may also be part of a functional gastrointestinal disorder (18).
Children younger than 8 years old, or even as old as 12 years, may
not be able to provide accurate descriptions of the degree, character,
and location of pain (4). Whether H pylori gastritis causes abdomi-
nal pain in the absence of PUD is still debatable. Several studies
from the 1990s applied different noninvasive tests for H pylori
JPGN
Volume 53, Number 2, August 2011
Recommendations for H pylori Infection in Children
www.jpgn.org
233
Copyright 2011 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
infection and compared the prevalence of positive results in
children with recurrent abdominal pain and controls and found
no significant difference in infection rates between cases and
controls (19,20). A meta-analysis of 45 studies concluded that
H pylori
infection is not associated with abdominal pain (21).
Epidemiological studies on the prevalence of chronic or recurrent
abdominal pain in pediatric age groups in different European
countries yielded estimated frequencies ranging from 0.3% to
19%; however, the frequencies in different countries were not
related to the background of H pylori prevalence in the respective
countries (4). More recent case-control studies confirmed the lack
of evidence for a causal relation between H pylori infection and
abdominal pain. In a study of 1221 children from Germany, Bode
et al (22) identified in a multivariable logistic regression analysis
that social and familial factors (single-parent household, family
history of PUD, or functional pain) were significantly associated
with abdominal pain, but not with the H pylori status of the child, as
assessed by the
13
C-UBT. Tindberg et al (23) reported no significant
association of recurrent abdominal pain with H pylori infection in
695 schoolchildren between 10 and 12 years old. In fact, an inverse
relation was noted for H pylori positivity and the occurrence of any
abdominal pain after adjustment for selected possible confounders
(odds ratio [OR] 0.5; 95% confidence interval [CI] 0.3–0.8).
Several uncontrolled intervention studies showed improve-
ment of symptoms after treatment; however, in some of the studies,
treatment success was not monitored and eradication of the bacteria
was assumed in cases with symptomatic improvement (12,24–26).
Other studies had a short follow-up period of a few weeks only (27).
These uncontrolled intervention studies provide weak evidence of
a causal relation between H pylori infection and abdominal pain,
particularly because functional abdominal pain resolves in 30% to
70% of patients by 2 to 8 weeks after diagnosis accompanied by
reassurance of the child and the parents (28,29).
Only 1 double-blind randomized placebo-controlled trial was
performed in a population of symptomatic children with H pylori
infection, excluding cases of PUD (30). In this small trial with 20
children studied for 12 months, a relation between symptom relief
and H pylori eradication or histological healing was not observed.
In summary, at present, there is inadequate evidence
supporting a causal relation between H pylori gastritis and
abdominal symptoms in the absence of ulcer disease. Therefore,
cases of abdominal pain consistent with the diagnostic criteria of
functional pain (18) should not be investigated for H pylori, unless
upper endoscopy is performed during the diagnostic workup in
search for organic disease.
Recommendation 3
In children with first-degree relatives
with gastric cancer, testing for H pylori may be considered.
Agree: 93% (AR 29%, A 50%, AS 14%, D 7%). Grade of
evidence: low.
Comment on Recommendation 3:
A causal relation between H pylori infection and the risk
of gastric malignancies, including cancer and gastric marginal zone
B-cell lymphoma of mucosa-associated lymphoid tissue (MALT)
type, has been shown in animal models and is supported by several
epidemiological and intervention studies (31–34). Both of these
cancer types are extremely rare during the first 2 decades of life.
Although H pylori–associated gastric cancer has not been reported
in children, MALT lymphomas have been described in a few
H pylori
–infected pediatric patients (2,3).
In 1994, the World Health Organization declared H pylori
a class I carcinogen. A meta-analysis estimated that the risk
for gastric cancer is increased by a factor of 2 to 3 in H pylori–
infected individuals. The risk is further increased if only noncardia
carcinomas are considered; however, the risk of gastric cancer not
only depends on the infection itself but also is strongly modified
by the presence of bacterial virulence factors (35) and other factors
such as the genetic makeup of the host and environmental
influences, including diet (36). The eradication of H pylori may
have the potential to decrease the risk of gastric cancer (37–39). In
a large interventional trial in adults, subgroup analysis suggested
that eradication may be beneficial in people without precancerous
lesions (39). The time point for an effective intervention,
and therefore screening strategy, however, is not yet clear (40).
In previous studies of patients younger than 45 years old with
gastric cancer, H pylori had been identified as a risk factor (41).
Individuals with a positive family history for gastric cancer
are considered a high-risk group. The risk may be particularly high
in H pylori–infected children in whom the father or the mother is
affected by gastric cancer. This child not only shares genetic and
environmental factors with the affected parent but may also have the
same bacterial strain with pathogenic properties (42,43). Therefore,
the risk of gastric cancer may be much higher for individual
children with such histories than what has been estimated from
epidemiological studies that lack information on relevant factors.
Although there is little evidence that addresses whether this
approach is beneficial, there was strong agreement within the panel
that testing for H pylori infection be considered in children with
a first-degree relative with gastric cancer. There was also agreement
that if H pylori infection is confirmed in these children either with
a reliable noninvasive test or with biopsy-based methods, treatment
should be offered and the success of therapy evaluated to ensure
successful eradication.
Approximately 70% of gastric MALT lymphomas can be
successfully treated with H pylori eradication. In the rare cases of
H pylori
–infected children with established MALT lymphoma,
eradication therapy needs to be performed regardless of the staging
of the lymphoma. The translocation t(11;18)(q21;q21) character-
istic of MALT lymphoma is recognized to be a marker of H pylori
independence, but this marker is found in only half of the MALT
lymphomas that are resistant to H pylori eradication (44). In patients
with the translocation t(11;18)(q21;q21), conventional chemo-
therapy can be considered in addition to eradication of H pylori.
Screening for H pylori infection in the general population is
not recommended. In populations with a high prevalence of H pylori
infection, the benefit of screening can be assessed by considering
the risk of H pylori–associated gastric cancer in particular popu-
lations, along with the health care priorities of those populations. In
populations with a high incidence of gastric cancer and in which
gastric cancer–screening programs are in place, children can be
included in screening programs for H pylori infection, and close
surveillance in those who develop atrophy or intestinal metaplasia is
indicated.
Recommendation 4
In children with refractory iron-defi-
ciency anemia in which other causes have been ruled out, testing
for H pylori infection may be considered.
Agree: 100% (AR 36%, A 36%, AS 28%). Grade of
evidence: low.
Comment on Recommendation 4:
Iron-deficiency anemia in children and adolescents may have
different causes. If noninvasive diagnostic tests are not able to find
the cause and/or if the iron deficiency is refractory to oral iron
therapy, then diagnostic upper endoscopy is indicated. In these
situations, mucosal biopsies are taken to rule out pathologic con-
ditions such as celiac disease. In addition, gastric biopsies are taken
for evaluation of H pylori by histology and culture because H pylori
infection may be the cause of iron-deficiency anemia, even in the
Koletzko et al
JPGN
Volume 53, Number 2, August 2011
234
www.jpgn.org
Copyright 2011 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
absence of erosions or ulceration (45,46) or gastrointestinal symp-
toms (47).
Several studies have shown an association between low iron
status and H pylori infection (48–50). Because both H pylori
infection and iron deficiency are associated with poor socioeco-
nomic and hygienic conditions, and cross-sectional studies cannot
determine whether the purported cause preceded the effect, only
randomized intervention studies can provide strong evidence of
a causal relation. The first randomized placebo-controlled study
included only 22 H pylori–infected pediatric patients randomized
into 3 treatment arms: iron only, eradication therapy only, or both
(48). Eradication therapy increased hemoglobin levels even without
iron substitution, whereas iron therapy alone did not. In a study
of 140 children between 6 and 16 years old from Turkey, it was
reported that eradication therapy in the absence of iron supple-
mentation was sufficient to improve iron deficiency and anemia
(49); however, this beneficial effect of H pylori therapy on iron
status could not be confirmed in recent intervention trials in children
living in Alaska and Bangladesh (50,50a). Further placebo-
controlled studies are needed to show whether H pylori infection
can cause iron deficiency even in the absence of mucosal breaks
because low iron status can have harmful effects on both mental and
physical development.
Recommendation 5
There is insufficient evidence that H pylori
infection is causally related to otitis media, upper respiratory
tract infections, periodontal disease, food allergy, SIDS,
idiopathic thrombocytopenic purpura, and short stature
.
Agree: 100% (AR 36%, A 28%, AS, 36%). Grade of
evidence: low.
Comment on Recommendation 5:
A wide variety of extraintestinal manifestations are
suggested to be associated with H pylori infection; however, current
evidence for a causal relation for these associations in children is
not compelling (51–62).
3.3. Which Diagnostic Test Should Be Applied
in Which Situation?
Numerous tests that detect H pylori are available. They are
divided into noninvasive and invasive tests. Invasive tests require
gastric tissue for detecting the organism and include culture, rapid
urease test, histopathology, polymerase chain reaction, and FISH.
(63). Noninvasive tests include different methods for the detection
of H pylori antigens in stool, detection of antibodies against
H pylori
in serum, urine, and oral samples, and the
13
C-UBT.
The sensitivities and specificities obtained in different pediatric
studies have been reviewed by the 4 members of the guideline
subgroup and recently published (63).
All diagnostic tests are generally feasible in children; how-
ever, tests requiring patient cooperation, such as the UBT, are more
difficult to perform in infants, toddlers, or physically challenged
children. A crucial question for all tests performed in a pediatric
population is whether the accuracy of the applied method is
influenced by the age of the tested child. It is necessary to consider
different age groups: infants, toddlers, preschool-age and school-
age children, and adolescents (64). Most of the validation studies in
children included only a few H pylori–infected infants and toddlers.
Therefore, the information with respect to sensitivity is limited in
these age groups.
It is necessary to compare a test to a reference standard;
however, no single test for detection of H pylori infection can be
used as a fully reliable reference method. Culture is the only method
that is considered to be 100% specific, a positive culture being
sufficient to prove H pylori infection, but its sensitivity is lower
(65,66). For that reason, concordant results of at least 2 tests are
needed to define the H pylori infection status. For noninvasive
tests, biopsy-based tests should be the reference. If culture was
not successful or not performed, concordant positive results for
histology and rapid urease test indicate a positive H pylori status.
The definition of a negative H pylori status is that all of 2 or
3 invasive tests performed are negative. For the validation of an
invasive test, such as histopathology, other biopsy-based tests, with
or without the combination of reliable noninvasive tests, should be
the reference. All of the tests are suitable for the detection
of infection before and after treatment, with the exception of
serology, which may remain positive for some time after successful
eradication.
For the interpretation of test results, factors that can lead to
false-positive or false-negative results must be known and
considered. Antibiotics, including penicillin and cephalosporines,
and acid-suppressive drugs, particularly PPIs, should be discon-
tinued before testing for at least 4 and 2 weeks, respectively. This
recommendation is extrapolated from adult studies (67–69).
Recommendation 6
For the diagnosis of H pylori infection
during EGD, it is recommended that gastric biopsies (antrum
and corpus) for histopathology are obtained.
Agree: 93% (AR 33%, A 40%, AS 20%, DS 7%). Grade
of evidence: moderate.
Recommendation 7
It is recommended that the initial
diagnosis of H pylori infection be based on either positive
histopathology R positive rapid urease test or a positive culture.
Agree: 100% (AR 36%, A 50%, AS 14%). Grade of
evidence: moderate.
Comment on Recommendations 6 and 7:
For histology, 2 biopsies should be obtained from both
the antrum and the corpus, and the findings should be reported
according to the updated Sydney classification (70). Because the
density of H pylori may be patchy, the sensitivity increases with the
number of biopsies taken. Normally, the highest bacterial count is
found in the antrum; however, in cases of low gastric acidity, the
bacteria may be present only in the corpus. In a small single-center
study of children undergoing endoscopy for symptoms of acid
peptic disease in Italy, in 22 children in whom H pylori infection
was identified, biopsies of the cardia were more sensitive for the
detection of H pylori than biopsies of the antrum or corpus (71);
however, these findings need to be confirmed in additional centers.
Special staining (Giemsa or silver stain) and immunohistochemistry
may improve the detection of H pylori. Biopsies should be stained
with hematoxylin and eosin for histopathology because this is the
best method to detect atrophy and intestinal metaplasia. Atrophy
can be assessed only in biopsy material that is oriented correctly,
and diagnostic concordance between pathologists can be difficult
to achieve. Histopathology also allows the recognition of the rare
Helicobacter heilmannii
infection (72).
In children with suspected H pylori infection, it is highly
recommended to take not only biopsies for histopathology but also
1 biopsy each for a rapid urease test and, if available, culture. The
suspicion of an infection is often based on the macroscopic findings
of a nodular mucosa in the antrum or bulbus and/or gastric or
duodenal erosions or ulcerations. The rationale for the recommen-
dation to perform more than 1 diagnostic test is based on the
sensitivity results of invasive tests, which range from 66% to
100% for histology and from 75% to 100% for rapid urease tests
in published series from children (63). With decreasing prevalence
JPGN
Volume 53, Number 2, August 2011
Recommendations for H pylori Infection in Children
www.jpgn.org
235
Copyright 2011 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
of the infection in pediatric populations in many areas of
Europe and North America, the predictive values of the diagnostic
test results fall. For example, a test with a sensitivity of 90% has a
positive predictive value of only 50%, if the prevalence of the
infection in the population is 10%. Therefore, concordant
positive results on 2 different tests are recommended to confirm
the diagnosis and justify the costs and adverse effects of treatment.
If the results of histology and rapid urease test are discordant, then
a noninvasive test (UBT or stool test) should be applied.
One exception from the rule of 2 concordant test results is a
positive culture, which is 100% specific and therefore in itself
sufficient to diagnose H pylori infection. Another exception is the
presence of a bleeding peptic ulcer, in which case 1 positive biopsy-
based test is considered to be sufficient to initiate anti-H pylori
therapy. A recent meta-analysis on the accuracy of diagnostic tests
in adults with PUD clearly indicated that active bleeding decreases
the sensitivity of invasive diagnostic tests, but the specificity is
high (73).
Recommendation 8
The
13
C-UBT is a reliable noninvasive test
to determine whether H pylori has been eradicated.
Agree: 94% (AR 67%, A 20%, AS 7%, DS 6%). Grade
of evidence: high.
Comment on Recommendation 8:
The UBT has been evaluated in a large number of pediatric
studies of high quality against a reference standard, both before
and after therapy (74 – 78). In spite of a high variability of tracer
dose and tracer application, the type of test meal, the duration of the
fasting period before the meal, the time point of breath sampling,
the type of analysis, and the cutoff levels, this test has a high
accuracy, sensitivity, and specificity (63,64). When the UBT is
performed, the patient should have an empty stomach before
receiving an acid drink (apple or orange juice, citric acid solution)
because the urease activity of the bacteria decreases rapidly
with increasing pH (79). After ingestion of the tracer, the drink
without tracer should be provided to the child to avoid degradation
of the tracer by oral flora. This is a particular problem in infants
and toddlers and may at least in part explain the lower specificity
reported in children younger than 6 years old compared with
older children (74,76,80 – 84). False-positive results can also
occur in young children because of the lower distribution
volume and a different CO
2
production rate, which can be adjusted
for (85).
Recommendation 9
A validated ELISA for detection of H
Dostları ilə paylaş: |