Gastrointestinal Manifestations in Children and Adolescents

Journal of Spirochetal and Tick-borne Diseases
The Spectrum of Gastrointestinal Manifestations in Children and
Adolescents
with Lyme Disease


Martin D. Fried, MD, Departments of Pediatric Gastroenterology; Matthew
Abel,
MD, Departments of Pediatric Gastroenterology; Dorothy Pietrucha, MD,
Pediatric
Neurology; Yen-Hong Kuo, MS, Academic Affairs; and Aswine Bal, MD,
Pediatric
Infectious Disease, Jersey Shore Medical Center, Neptune, New Jersey


Abstract
A clinical diagnosis of Lyme disease was made in 15 consecutive
patients
between the ages of 8 and 20 years who presented with a history of an
erythema
migrans rash followed by chronic gastrointestinal symptoms and multiple
organ
system complaints. Endoscopic evaluation was performed to assess the
gastrointestinal mucosa and to obtain biopsies for polymerase chain
reaction
(PCR) to the outer surface protein A (Osp A) of Borrelia burgdorferi.
As age
matched controls, 10 patients with biopsy-proven Crohn's disease were
also
tested by PCR. The laboratories assessing the histopathology and
performing
the
PCR were blinded to the diagnosis of all specimens.
The presence of B burgdorferi DNA in the gastrointestinal tract was
confirmed
by PCR in all of the patients with the clinical diagnosis of Lyme
disease who
had chronic gastrointestinal symptoms and in two control subjects with
Crohn's
disease. Biopsy evidence of chronic gastritis, chronic duodenitis, and
chronic
colitis was found in patients with Lyme disease who had chronic
gastrointestinal symptoms and was associated with the presence of B
burgdorferi.


The chronic gastrointestinal symptoms that occurred within 6 months of
an
erythema migrans rash and Lyme disease may be attributed to a direct
effect
or
immune mediated response to B burgdorferi. [J Spiro Tick Diseases
6(4):89-93,
1999. © 1999 Lyme Disease Foundation,


Introduction
Lyme disease affects a wide range of organ systems, producing
dermatologic,
musculoskeletal, neurologic, genitourinary, lymphatic, hepatic, renal,
respiratory, cardiovascular, and ocular manifestations.[1,2] One report
to
date
describes the presence of Borrelia burgdorferi in the stomach,
intestines,
and
colon of children.[3] To further address the clinical manifestations of
Lyme
disease and the possibility of direct involvement of the
gastrointestinal
(GI)
tract, a prospective study was made of 15 consecutive patients who had
a
physician documented erythema migrans (EM) rash followed by chronic
gastrointestinal symptoms and multiple organ system complaints of Lyme
disease..


Methods
All patients included in our study had a physician documented EM rash
with no
prior history of gastrointestinal complaints. They were referred to the

pediatric gastroenterology and nutrition service of Jersey Shore
Medical
Center
for evaluation of chronic abdominal pain, chronic diarrhea, acid
reflux, or
blood in the stool that occurred within 6 months after the onset of the
EM
rash. From January 1998 through April 1999, 15 consecutive patients
satisfying
the above clinical criteria[4] were evaluated prospectively. There were
6
boys
and 9 girls evaluated (mean age 1463.6 years, range 8-20). Each case
included
a
history, physical examination, complete blood cell count, liver
function
tests,
sedimentation rate, antinuclear antigen (ANA), HLA B27,
esophagogastroduodenoscopy (EGD), and/or colonoscopy. A Lyme Western
blot was
performed for confirmation of an acute (immunoglobulin M) or past
(immunoglobulin G) B burgdorferi infection. A positive IgM Western blot
was
interpreted as 2 of 3 bands (23, 39, 41 kd). A positive IgG Western
blot was
interpreted as 5 or more of the following B burgdorferi-specific bands:
18,
23,
28, 31, 34, 39, 41, 45, 58, 66, 93 kd. A diet history was taken to
assess the
dietary fat intake. Ultrasonography of the abdomen was performed when
the
history suggested a diagnosis of gallstones or pancreatitis. Stool
samples
were
examined for occult blood, Salmonella, Shigella, Yersinia,
Campylobacter, ova
and parasites, and Clostridium difficile toxin. Gastrointestinal (GI)
biopsies
were reviewed to assess the mucosa by microscopy and whether
Helicobacter
pylori (on EGD only) or eosinophilia was present.
Biopsy specimens were taken from areas of the GI tract that looked
inflamed
during EGD or colonoscopy. The biopsies were assigned randomly to three

histopathologists who were blinded to the diagnosis of the specimens
they
received. The histopathologists did not perform a silver stain for the
detection of spirochetes because it is not routinely done. Biopsies
were
reported as acutely inflamed when polymorphonuclear cells were present
in the
mucosa and chronically inflamed if 6 or more plasma cells and
lymphocytes
were
present in the gastric mucosa without polymorphonuclear cells. Chronic
duodenitis or chronic colitis was diagnosed when more than 6
intraepithelial
lymphocytes per 100 surface absorptive cells were present in tissue
biopsies
in
conjunction with a distortion in glandular architecture.


Polymerase chain reaction (PCR) for DNA to B burgdorferi[5] was
performed on
all biopsies by Medical Diagnostic Laboratories in Mount Laurel, New
Jersey.
In
all patients in which B burgdorferi DNA was detected, PCR for B
burgdorferi
RNA
polymerase was performed and results are reported in the Table. As a
target
for
DNA amplification, the gene coding for the outer surface protein A
(OspA) of
B
burgdorferi was selected and analyzed as described below.


DNA Isolation from Biopsy Specimens


Total DNA was extracted from duodenal, gastric, and colonic biopsies as

described by Maniaties et al.[6] The samples were centrifuged (5
minutes,
4°C,
14K rpm) and the pelleted biopsy was subjected to 500 µL of cell lysis
buffer
[0.5% SDS, 470 µL TE buffer, 5 µL of proteinase K (20 µg/µL)]. The
samples
were
incubated for 24 hours at 50°C. Proteinase K (5 µL) was added to the
mixture
every 6 hours. DNA was extracted by phenol chloroform, followed by
ethanol
precipitation. DNA concentrations were determined
spectrophotometrically by
measuring the A260.


DNA Amplification


The SL primers (SLA 59-AAT AGG TCT AAT AAT AGC CTT AAT AGC-39 SLB 59
CTA GTG
TTT TGC CAT CTT CTT TGA AAA-39) are suitable for amplification of all B

burgdorferi sensu lato isolates. The SL primers amplify a region
(nucleotide
21-328) of the B burgdorferi senso stricto B31 OspA sequence. One µg
of
isolated DNA was used as a template DNA in the presence of a 20 pmol
sample
of
each primer in a 50 µL reaction mixture. The samples were subjected to
35
amplification cycles in a Perkin Elmer 2400 thermocycler (Foster City,
CA)
under the following conditions: 93°C, 1 minute; 65°C, 1 minute; and
72°C, 1
minute. PCR amplification products were resolved onto 1.5% agarose
electrophoresis gels and visualized under ultraviolet light with
ethidium
bromide.


To test the presence of inhibitory substances and to provide a positive

control
in the PCR assay, amplifications were also performed with primers
targeting
the
histone gene. A positive control was performed with every biopsy
specimen. It
included a PCR in the presence of 100% B burgdorferi DNA that was
purchased
from the American Type Culture Collection (Rockville, MD). This B
burgdorferi
DNA was isolated from Ixodes scapularis tick, New York Type strain, and

shipped
frozen to the laboratory.[7] The negative control performed with each
biopsy
specimen included the PCR in the absence of DNA. A second genomic DNA
control
is done weekly at the laboratory as part of their quality control.
Physical
containment measures ensured the absence of DNA contamination in the
PCR
procedure.


As age-matched controls, 10 adolescents with biopsy proven Crohn's
disease (5
boys, 5 girls, 13.562.5 years, range 10-17), who had not been on
antibiotics
one year prior to endoscopy, were also tested by PCR. The laboratory
performing
the PCR analysis was blinded to the diagnosis of all specimens they
received.


Statistical Analysis


The sensitivity and specificity of PCR for the detection of B
burgdorferi in
the GI tract was calculated. The confidence intervals (CI) were
calculated by
using the Fischer's Exact test method. A Fischer's exact test was used
to
determine the association between inflammation and PCR positivity in
each of
the biopsied site.


Results
Patients with Lyme disease presented with chronic abdominal pain (n=10,
67%),
chronic diarrhea (n=1, 7%), visibly evident blood in the stool (n=2,
13%),
and
acid reflux with heartburn (n=2, 13%). In all 4 patients whose biopsies

revealed evidence of colitis, the abdominal pain was characterized as a

crampy,
periumbilical pain that started at the right middle quadrant of the
abdomen
and
spread to the left middle quadrant of the abdomen or vice versa. The
pain was
unrelated to meals and occurred throughout the day. In the remaining 6
patients
with abdominal pain whose biopsies revealed gastritis, duodenitis, or
both,
the
abdominal pain was characterized as periumbilical, burning, and
improved by
avoiding fried foods and foods high in fat content. Ultrasonography of
the
abdomen did not reveal any gallstones or evidence of pancreatitis. In 2

patients who complained of acid reflux, their pain was a burning
midepigastric
pain that radiated to the esophagus. The pain occurred within the first

postprandial hour and was relieved by antacids. Ten of the 15 patients
with
Lyme disease had evidence of inflammation at a biopsy site with
detection of
B
burgdorferi DNA at that site. Patients 2 and 11 had blood in their
stool and
presented with the clinical features of Crohn's disease (ie, 15 pound
weight
loss in a year, arthritis of the knee, protein losing enteropathy) and
ulcerative colitis (6 bloody bowel movements a day for a week),
respectively.
The biopsies of all the patients with Lyme disease revealed no evidence
of
granulomas or terminal ileitis. In patient 6, the IgM Western blot was
positive
and showed the 23, 31, 34, 39, 41, 58, and 66 kd bands. The IgG Western
blot
was negative (no bands present). No other patient had a positive
Western
blot.
All control patients with Crohn's disease had biopsy proven terminal
ileitis
and granulomatous colitis. Lyme disease was diagnosed in 15 patients
and 2
with
Crohn's disease had a positive PCR to B burgdorferi DNA in biopsy
specimens
from the gastrointestinal tract (Table). In 6 patients with Lyme
disease, B
burgdorferi DNA was detected in the GI tract and B burgdorferi RNA
polymerase
was detected by PCR.
A positive B burgdorferi PCR occurred with chronic inflammation in the
GI
tract
of 11 of 15 patients with Lyme disease. In patients 8 and 9,
inflammation
occurred in the stomach; however, B burgdorferi DNA was detected in the

colon.
B burgdorferi DNA was detected in the GI tract in the absence of
inflammation
in 4 patients (27%), 3 of whom had received at least 2 months of
antibiotics
prior to endoscopy. There was no statistically significant association
between
PCR positivity in the GI tract and chronic inflammation.


In 10 of 15 patients (67%), antibiotic therapy for Lyme disease had
been
prescribed within 1 to 5 months prior to endoscopy (n=4, 1-2 months;
n=3, 3-4
months, and n=3, 5 months). Despite prior antibiotic use, all 4
patients with
colitis were PCR positive for B burgdorferi DNA in the colon while 5 of
9
with
gastric inflammation were PCR positive in gastric biopsies.
Helicobacter
pylori
was not detected in any of the gastric biopsies. Salmonella, Shigella,
Yersinia, Campylobacter, and Clostridium difficile toxin was not
detected in
any of the stool samples. HLA B27 was positive in patients 1 and 11 and
in
none
of the controls. ANA was positive and had a speckled pattern in
patients 2,
5,
and 6. An elevated sedimentation rate of 85 and 28 were found in
patients 4
and
13, respectively.


The lab performing the PCR had a false positive rate of 1 in 500 by
analyzing
6550 specimens from January 1998 through April 1999. The sensitivity of
GI B
burgdorferi DNA detection was 100% (15/15) with a 95% CI (81.9%, 100%).
The
specificity was 80% (8/10) with a 95% CI (44.4%, 97.5%). The positive
predictive value was 88.2% (15/17) with a 95% CI (63.6%, 98.5%).


The Spectrum of Gastrointestinal Manifestations continued...


Conclusion
Children and adolescents with a history of Lyme disease and chronic GI
symptoms
occurring within 6 months of an EM rash, had evidence of inflammation
in the
stomach, duodenum, and colon. We found B burgdorferi by PCR of GI
biopsies to
be associated with chronic inflammation. The inflammatory reaction we
describe
may have been caused by spirochetes or by immune system products
elicited in
response to the spirochete presence.

.

Relevant Pages