264 Hong Kong Med J Vol 12 No 4 August 2006
Intestinal tuberculosis in a regional
hospital in Hong Kong: a 10-year
Hong Kong Med J 2006;12:264-71
United Christian Hospital, Kwun Tong,
Department of Medicine and Geriatrics
VKS Leung, FRCP (Lond, Edin, Glasg), FHKAM
ST Law, MRCP (UK)
TN Chau, MRCP (UK), FHKAM (Medicine)
WH Chan, MRCP (UK)
Department of Surgery
CW Lam, FRCS (Glasg), FHKAM (Surgery)
SH Lam, FRCS (Glasg), FHKAM (Surgery)
Department of Pathology
ISC Luk, FRCPath, FHKAM (Pathology)
Department of Radiology
TKL Loke, FRCR, FHKAM (Radiology)
Correspondence to: Dr VKS Leung
VKS Leung ??
SH Lam ??
ST Law ??
CW Lam ??
ISC Luk ??
TN Chau ??
TKL Loke ??
WH Chan ??
Objective. To study the clinical and pathological characteristics of patients with
Design. Retrospective study.
Setting. United Christian Hospital, Hong Kong.
Patients. Patients with intestinal tuberculosis diagnosed between January 1995
and December 2004 inclusive.
Results. The median age of the 13 male and 9 female patients was 53 years
(range, 12-81 years). Nineteen (86%) had a definitive diagnosis of intestinal
tuberculosis confirmed by the presence of caseating granulomas and/or acid-fast
bacilli in histological specimens. In three (14%) the diagnosis was based on
histology revealing non-caseating granulomas and a positive response to
anti-tuberculous treatment. Common symptoms included abdominal pain (82%),
diarrhoea (55%), weight loss (55%), and fever (45%). Three (14%) of the
patients were complicated by intestinal obstruction, and another two (9%) had
intestinal perforation. Four (18%) had concomitant active pulmonary tuberculosis.
The most frequently involved site was the ileocaecal region, which was affected
in 19 (86%) patients. Other sites included the jejunum, ascending and sigmoid
colon. The diagnosis of intestinal tuberculosis was facilitated by examination of
colonoscopic biopsy specimens (11 patients), and by examination of resected
surgical specimens in the remainder. Two patients died from terminal malignancy.
The remainder completed anti-tuberculous therapy and responded satisfactorily.
Conclusions. The diagnosis of intestinal tuberculosis is difficult due to the lack
of specific signs or symptoms. Colonoscopy with ileoscopy are useful tools in
the search for colonic and terminal ileal tuberculosis. Surgical exploration is
reserved for equivocal cases and for those who present as emergencies.
?? NP?? !"V?? !"#$%&'(RP??? NO?UN???NV?
NN?? !"#$%&'()*+,-.! /0123?4?56789:(
?? ?? !"##$%&'(#)*+,
??P? ?NQB? ?? !"#$%&'()*+,-./0123456789
?? ?? !"#$%&'()*$+,-./01234(5$)(5678
?? ?? !"
?? !"?#$%&' ?UOB? ?? ?RRB? ?? !" ?RRB? ?? ?QRB? ?
?? ?? !"#$%&'
P??NQB??? !"#$O??VB??? !"Q??NUB??? !"#$
?? NVVR ?N??OMMQ?NO?? !"#$%&'()*)+,
?? !"#$%&'?NV??USB??? !"#$%&'(%)*+(%,
?USB??? !"#$%&'()*+,-L?? !"#$%&'()*+,
?? !"O?? !"#$%&'()*?+,-./01123456789
In recent years, the incidence of intestinal tuberculosis (TB) in developed
countries has been increasing along with the overall resurgence of TB.1,2 There
were an estimated 8.3 million new TB patients globally in 2000; newly
Hong Kong Med J Vol 12 No 4 August 2006 265
Intestinal tuberculosis in Hong Kong
diagnosed cases increased at a rate of 1.8% per year
between 1997 and 2000.3 This resurgence is, in part, due to
the pandemic of human immunodeficiency virus (HIV)
infection.1-3 The overall incidence of TB in Hong Kong has
remained high with about 90 cases per 100 000 inhabitants
in the year 2005,4 but details pertaining to intestinal TB
are not available.5 The clinical manifestations of intestinal
TB are non-specific; a high index of suspicion is therefore
important to ensure a timely diagnosis; missed or delayed
recognition can result in avoidable morbidity and
mortality.1,2 The purpose of this study was to retrospectively
evaluate the clinical and pathological characteristics,
diagnostic modalities, and treatment response of patients
with intestinal TB.
Patients who presented to the United Christian Hospital
(a regional hospital in Kowloon East serving a population
of over 600 000) between January 1995 and December
2004 inclusive diagnosed to have gastro-intestinal (GI) TB
were identified. Their medical records, chest radiographs
(CXR), and histological sections were reviewed. Patients
were classified as having either confirmed or presumed
intestinal TB. The diagnosis of confirmed TB was based
on affected tissues with one or more of the following:
(1) histological evidence of granulomas with caseating
necrosis, (2) demonstration of acid-fast bacilli (AFB)
by Ziehl-Neelsen (ZN) stain, or (3) culture of Mycobacte-
rium tuberculosis.6 Patients in whom the diagnosis was
suspected though unconfirmed by histology or microbiol-
ogy were regarded as having presumed intestinal TB, if
they were treated successfully with anti-tuberculous
chemotherapy. A telephone survey was conducted on all
patients to confirm completion of anti-tuberculous treatment
and resolution of symptoms.
Over the 10-year period, 58 patients at our institution were
diagnosed with abdominal TB. Sites of infection were: the
peritoneum (n=34; 59%), the GI tract (n=22; 38%) and in
the liver (n=1) and mesenteric lymph nodes (n=1). Table 1
lists the numbers of all TB patients and those with intestinal
TB diagnosed at our institution according to the year of
diagnosis. Six intestinal TB patients presented during the
first 5 years of the study, and 16 during the last 5 years.
All 22 patients with intestinal TB were of Chinese ethnicity;
13 were male and 9 female. Their median age was 53
(range, 12-81) years; 54 years for the six patients
presenting during the first 5 years and 53 years for the 16
Seven (32%) of the patients had a history of prior
treatment for pulmonary TB, and one for tuberculous
cervical lymphadenitis. Four (18%) had active pulmonary
TB when intestinal TB was diagnosed, and one of them also
had a perianal abscess.
Only one patient with intestinal TB was in receipt of
immunosuppressive agents (prednisolone and azathioprine
for the treatment of optic neuritis and transverse myelitis).
This patient also had a pelvic liposarcoma and post-
radiation proctosigmoiditis. Other coexisting medical
illnesses include diabetes mellitus in three patients,
previous stroke in two, and advanced colonic adenocarci-
noma in two. Sixteen (73%) of the patients had no
coexisting medical illness. One patient reported
consumption of unpasteurised milk prior to the development
of abdominal symptoms.
Eleven (50%) of the patients with intestinal TB presented
insidiously, while nine (41%) presented abruptly, six of
whom had an acute-on-chronic presentation. The median
duration of symptoms was 1.5 months (range, 0.03-12
months). Two patients with advanced adenocarcinoma of
the colon (one in the caecum, another in the sigmoid
colon) were diagnosed to have coincidental mycobacterial
infection in the resected tumour specimens.
The presenting symptoms of intestinal TB are listed in
Table 2, the most frequent being abdominal pain (82%).
Others included diarrhoea, constipation, nausea and
Table 1. Number of patients with (i) tuberculosis (TB) and (ii) intestinal TB attending the Tuberculosis and Chest Clinics and at
the United Christian Hospital (UCH)
Year No. of TB
No. of new TB cases
attending TB and
No. of new intestinal
TB cases attending
TB and Chest Clinics11
No. of TB
No. of new cases
of intestinal TB
Leung et al
266 Hong Kong Med J Vol 12 No 4 August 2006
vomiting, and rectal bleeding. About half of them had
fever, anorexia, and weight loss. Respiratory symptoms
were noted in the four patients with concomitant active
Four (18%) patients presented with acute right lower
quadrant pain masquerading as acute appendicitis. Three
(14%) presented with intestinal obstruction, which resolved
after conservative treatment and colonoscopy, except in
one patient who was subjected to exploratory laparotomy.
Two (9%) patients presented with intestinal perforation and
one developed perforation at the terminal ileum 3.5 months
after starting anti-tuberculous chemotherapy for pulmonary
and perianal TB.
Laboratory and radiological abnormalities
Ten (45%) patients had anaemia, of whom two also had
colon cancer. Leukocytosis was present in 10 (45%)
patients, the erythrocyte sedimentation rate was elevated
in nine (41%), and eight (36%) had hypoalbuminaemia.
All patients with an abrupt-onset presentation had one or
more abnormal laboratory parameters detected, whereas
eight of the 11 patients who presented insidiously had no
abnormalities detected in routine blood tests. Test for HIV
antibodies was performed in nine individuals and all were
Eight (36%) patients had radiological evidence of
prior pulmonary TB, and four (18%) of active pulmonary
TB. Only one of these patients underwent abdominal
radiological imaging (computed tomographic scan and a
small bowel enema), which revealed mucosal irregularities
and ulcerations in the terminal ileum.
Sites of involvement
The frequencies of different sites of involvement along the
GI tract and the colonoscopic features that were noted are
shown in Table 3. Most of the patients (86%) had ileocaecal
involvement. Skip lesions along the intestine were described
in three (14%) patients—two had involvement of terminal
ileum and ascending colon, and one involved the caecum
and ascending colon.
Diagnosis of intestinal tuberculosis
The diagnosis of intestinal TB was made by examination of
biopsy specimens obtained at colonoscopy (n=11), and
surgical specimens obtained at laparotomy (n=11).
Intestinal TB was suspected before colonoscopy in only
four patients—three of whom had concomitant active
pulmonary TB, and in the fourth patient terminal ileal
ulcerations had been detected by radiological imaging. None
of the patients who underwent surgery were suspected to
have intestinal TB preoperatively.
Nineteen (86%) patients had confirmed intestinal TB
(positive histology alone in 2, positive histology and
positive ZN stain for AFB in 1, positive stain for AFB
only in 13, positive stain for AFB plus positive culture for
mycobacteria in 2, and positive culture for mycobacteria
alone in 1). The diagnosis of intestinal TB could not
be confirmed in three (14%) patients who eventually
responded to anti-tuberculous treatment. One of the
latter had concurrent active pulmonary TB, and had
already commenced anti-tuberculous treatment at the
time of colonoscopy. A 12-year-old boy, who had
consumed unpasteurised milk and had terminal ileal
ulcerations revealed by radiological imaging, had
also started empirical anti-tuberculous treatment at
the time of colonoscopy. All three patients were
presumed to have intestinal TB and after anti-tuberculous
treatment they experienced symptom resolution and
disappearance of mucosal abnormalities on follow-up
Only two of the six cases of intestinal TB identified
during the first 5 years of the study were diagnosed before
the onset of severe symptoms or complications; in one of
them the diagnosis was facilitated by colonoscopy. Whereas,
during the latter half of the study period, 11 (69%) of 16
cases were diagnosed in the early stages of infection, and
colonoscopy helped to confirm the diagnosis in nine of these
Colonoscopy was attempted in 13 patients, in whom full
colonoscopy with successful terminal ileal intubation was
achieved in 12 cases. In one patient, colonoscopy was
incomplete as advancement of the endoscope was limited
Table 2. Presenting symptoms among patients with intestinal
Nausea and vomiting
Cough and sputum
No. of patients (n=22)
Table 3. Sites of involvement and colonoscopic features of
No. of patients
Colonoscopic features (n=12)
Deformed ileocaecal valve
Hong Kong Med J Vol 12 No 4 August 2006 267
Intestinal tuberculosis in Hong Kong
by severe histology-confirmed post-radiation colitis
affecting the rectosigmoid colon. The endoscopic features
noted in these 12 cases of complete colonoscopy are listed
in Table 3. Ulceration was present in all cases; most ulcers
being discrete and ranged from 0.3 to 2.5 cm in diameter.
One patient had a large circumferential ulcer in the terminal
ileum (Fig a). Ulcers with nodular edges (Fig b) were found
in five patients; four individuals had a deformed ileocaecal
valve (Fig c), two had mucosal nodules (Fig d), and one
had polypoid lesions. In four (33%) patients, terminal ileal
ulcers were the only abnormalities encountered at
endoscopy. In all 12 patients undergoing colonoscopy,
biopsy specimens were obtained from the site of the
abnormal mucosa, of which all were sent for histological
examination and seven for mycobacterial culture.
Operative findings and management
Exploratory laparotomy was carried out in 11 patients, of
which two were for symptoms attributed to complications
from cancer of the colon. Among the other nine, four had a
working diagnosis of acute appendicitis, two were believed
to have perforated peptic ulcers, and two were suspected to
have intestinal obstruction (from recurrence of a pelvic
liposarcoma, and a malignant caecal tumour). Three
patients were found to have an ileocaecal mass, all of whom
underwent right hemicolectomy with primary ileocolonic
Fig. Colonoscopic appearance in patients with intestinal tuberculosis
(a) a circumferential ulcer in the terminal ileum, (b) a discrete ulcer with nodular edges in terminal ileum (arrow), (c) an ulcerated and
deformed ileocaecal valve (arrow), and (d) mucosal nodules in the ascending colon
Leung et al
268 Hong Kong Med J Vol 12 No 4 August 2006
anastomosis. Two patients were found to have solitary
perforations in the terminal ileum; right hemicolectomy with
primary ileocolonic anastomosis was performed for one and
segmental resection of the terminal ileum with primary
anastomosis for the other. Two others had an ulcerative
caecal mass; one was treated by right hemicolectomy with
primary anastomosis, and the other only had excisional
biopsies of the enlarged mesenteric lymph nodes. One
patient with an inflamed and ulcerated segment of jejunum
(identified at surgery), had the affected bowel resected and
a primary anastomosis performed. Another had multiple
transverse ulcers in the terminal ileum in association with
a fibrotic and stenosed ileocaecal valve, resulting in intesti-
nal obstruction. This patient had a right hemicolectomy with
primary anastomosis to relieve the obstruction. Mesenteric
lymphadenopathy was present in nine patients (including
the two with colon cancer), and all underwent excisional
biopsies to establish a tissue diagnosis. Other abnormalities
included fibrous adhesions of bowel loops in four patients,
and visceral and parietal peritoneal nodules in one. In three
patients, tuberculous infection was suspected at the time of
operation and only the surgical specimens of these patients
were sent for culture of mycobacteria.
Pathology and microbiology
Histological and microbiological features in the
colonoscopic biopsy and surgical specimens are
summarised in Table 4. Granulomas were identified in 11
of 12 colonoscopic biopsy specimens, but caseating
granulomas in only one patient and AFB were identified by
ZN stain in seven. Three colonoscopic biopsy specimens
had a positive yield for culture of mycobacteria, one of
which had been negative for AFB with ZN stain. Among
the three patients whose tissue cultures grew mycobacteria,
one also had positive cultures from sputum and gastric
aspirates, and another had positive cultures from sputum
and bronchial aspirates. One patient who had unrevealing
histology in colonoscopic biopsy specimens and eventually
underwent exploratory laparotomy.
Granulomas were demonstrated in all surgical specimens
(in the lymph nodes of the nine patients with mesenteric
lymphadenopathy, and in the peritoneal nodules in one
other), and in seven of these there was caseating necrosis.
Caseating granulomas were found in the intestinal
pathology and mesenteric lymph nodes in four cases, and
were present only in the mesenteric lymph nodes in three.
Ziehl-Neelsen stain demonstrated AFB in nine cases, and
in another case in which the stain for AFB was negative,
mycobacteria were nevertheless cultured.
All four isolates of mycobacteria from intestinal tissues
were of the mycobacterium TB complex, and further
identification of the specific species was performed for three
isolates, which demonstrated they were all M tuberculosis.
All isolates were susceptible to isoniazid, rifampicin,
ethambutol, and streptomycin.
Chemotherapy and treatment outcome
Three patients were already on anti-tuberculous treatment
at the time intestinal TB was diagnosed. The remaining 19
patients commenced anti-tuberculous drugs after being
diagnosed—18 received standard treatment (rifampicin,
isoniazid, pyrazinamide, and ethambutol in the initial
2 months followed by rifampicin and isoniazid in the
continuation phase). One elderly patient commenced a
three-drug regimen consisting of rifampicin, isoniazid, and
pyrazinamide, because monitoring for ethambutol-related
visual impairment appeared unreliable. Seventeen of
these 22 patients were referred to TB and Chest Clinics
operated by the Department of Health of Hong Kong for
continuation of anti-tuberculous therapy. Twenty patients
completed their anti-tuberculous treatment (mean duration,
9.8 months; range, 6-12 months). All 22 patients
enjoyed resolution of abdominal symptoms after
anti-tuberculous treatment and there was no mortality
from TB, but two who had colon cancers died from their
terminal malignancy before completion of their treatment
Mycobacterium tuberculosis, along with Mycobacterium
bovis, Mycobacterium africanum, and Mycobacterium
microti, cause TB and are members of the mycobacterium
TB complex. In Hong Kong, M tuberculosis is the species
most commonly associated with mycobacterial infection.5,7
Mycobacterium tuberculosis is also the most common
species causing TB of the GI tract.8 Mycobacterium bovis is
a major cause of TB in cattle, and can be transmitted to
humans through consumption of unpasteurised milk
products from infected cows resulting in GI tract infection.
Nowadays, M bovis is an uncommon human pathogen in
Table 4. Histological and microbiological features of intestinal tuberculosis (ITB)
Colonoscopic biopsy specimens (n=12)
ITB (n=8)ITB (n=3)
Surgical specimens (n=11)
Positive Ziehl-Neelsen stain for acid-fast bacilli
Positive tissue culture for mycobacteria
* This patient subsequently underwent laparotomy
†One patient had positive culture of mycobacteria from sputum
Hong Kong Med J Vol 12 No 4 August 2006 269
Intestinal tuberculosis in Hong Kong
Intestinal TB was common in the early 20th century.
Old postmortem reports showed that up to 80% of patients
who had died from pulmonary TB also had intestinal
involvement.8 However, by the mid-20th century, all
forms of TB had declined dramatically in developed
countries as a result of increased standards of living,
control of bovine TB through slaughter of infected cattle
herds, pasteurisation of milk, and the introduction of anti-
tuberculous chemotherapy.2 In recent years, the incidence
of intestinal TB has increased again along with an
overall resurgence of TB.1,2 This resurgence is probably
related to the pandemic of HIV infection, increased use of
immunosuppressive drugs, and an ageing population.1,2,9,10
There are no detailed data on the overall number of
GI TB cases in Hong Kong, but it appears that the gut is
uncommonly affected, despite the high number of all TB
cases. Tuberculosis of the GI tract accounted for only 31
(0.74%) of the 4186 new TB cases attending TB and Chest
Clinics in Hong Kong in 2004 and accounted for less
than 2% of the 286 TB-related deaths in the same year.11
Nevertheless, in recent years the number of new cases of
intestinal TB attending TB and Chest Clinics appears to be
on the rise (Table 1).11 The reason for this increase is not
apparent, since there has been no parallel increase in all
cases of TB during the corresponding period (Table 1),4 and
in Hong Kong the prevalence of HIV-positive patients
with TB remains low.5 Our institution too has encountered
an increasing number of intestinal TB cases in recent
years, though none of our patients were HIV positive. Nor
did immunosuppressive drug use or ageing appear to be
major contributing factors. Although we encountered only
a small number of intestinal TB cases, mostly they were
diagnosed in the early phases of the infection, before the
development of severe symptoms or complications. The
increased use of colonoscopy as a tool for evaluation of
abdominal symptoms may have facilitated early diagnosis
of intestinal TB, and may, at least in part, explain the
increased number of cases of intestinal TB in recent years.
The pathophysiology of intestinal TB has been
attributed to four mechanisms: (1) haematogenous spread
from active pulmonary or miliary TB, (2) swallowing of
infected sputum in patients with active pulmonary TB, (3)
ingestion of milk or food contaminated with M bovis, and
(4) contiguous spread from adjacent organs.2 In one of
our patients with concomitant active pulmonary TB,
growth of mycobacteria from gastric aspirate suggests
swallowed infected sputum as the source of the infection.
Mycobacterium bovis infection was suspected in one
patient, who had consumed unpasteurised milk before the
onset of abdominal symptoms.
The diagnosis of intestinal TB is notoriously difficult,
as its presenting signs and symptoms and laboratory
abnormalities are non-specific and non-diagnostic.1,2,8-10,12
As illustrated in our cohort, routine blood tests being
normal cannot exclude intestinal TB. An important clue is
the presence of concomitant active pulmonary TB, but
this was only present in about 15% to 25% of patients.1,2,10
Intestinal obstruction is the most common complication
and occurs in about 20% to 30% of patients.1,9,10,12 Free
intestinal perforations ensue in 1% to 15% of cases,13-15 and
may occur after the commencement of anti-tuberculous
treatment.16-18 In one report, paradoxically, a patient
developed intestinal perforation 3.5 months after the
initiation of anti-tuberculous treatment.19
Evidently, the most frequent site of GI TB is the
ileocaecal region, where over 50% of the lesions are
located.1,2,8,10,12 This is also consistent with our observations
in this study. Predilection of the bacillus for the ileocaecum
is attributed to the abundance of lymphoid tissue in
that region, the relative physiologic stasis and minimal
digestive activity of the area causing increased contact
time between the bacteria and the intestinal mucosa, as well
as the high rate of absorption in the region.2,8,10,12 Other
commonly involved sites are the colon and the jejunum,
whereas the oesophagus, stomach, duodenum, and anus are
rarely involved.1,8,10,12 Anal TB may present with ulcers,
fistulae, abscesses, and as hypertrophic growths.8,12 One of
our patients also had a concomitant perianal tuberculous
abscess, which could have been a complication of a
tuberculous anal fistula, although we did not search for a
fistula in this case. Concomitant tuberculous peritonitis, as
seen in one patient, is encountered occasionally and related
to contiguous spread from tuberculous lesions of the
intestine.1 Involvement of regional mesenteric lymph nodes,
nonetheless, is common.8
Confirmation of the diagnosis of intestinal TB requires
isolation of M tuberculosis from affected tissue or evidence
of TB elsewhere with caseating granulomas in intestinal
tissue.8 In 1950, Hoon et al20 proposed the following criteria:
(1) growth of M tuberculosis from affected tissue, or (2)
histological demonstration of mycobacteria in tissue, or
(3) histological evidence of granulomas with caseating
necrosis in tissue, or (4) typical gross pathological findings
in the bowel and histological findings of granulomas with
caseating necrosis in associated lymph nodes.
Traditionally, surgical exploration was used to obtain a
tissue diagnosis. Although surgery may still be required in
patients who present as emergencies, many cases of GI TB
can now be diagnosed endoscopically. Colonoscopy with
procurement of biopsy specimens is currently considered
the most valuable diagnostic tool for identifying such
lesions in the colon and terminal ileum.21-26 Mucosal ulcers
and nodules are the most commonly encountered endoscopic
lesions, being present in over 80% to 90% of cases,23-26 and
in about half the ileocaecal valve is deformed.24,25 Other
endoscopic findings include: strictures, polypoid lesions, and
fibrous strands.23-26 Tuberculous ulcers are characteristically
aligned transversely along the intestine or are circumferen-
tial,8,10,12,26 however none of these endoscopic features are
diagnostic. Therefore endoscopic biopsy specimens should
Leung et al
270 Hong Kong Med J Vol 12 No 4 August 2006
always be examined histologically for granulomas and
AFB, and cultured for mycobacteria.1,2,6,8,25 Four of our
patients who underwent colonoscopy only had lesions in
the terminal ileum, highlighting the importance of perform-
ing ileal intubation for the diagnosis of intestinal TB.27
The most pathognomonic histological lesions of
intestinal TB are caseating granulomas, but their presence
in colonoscopic biopsy specimens is highly variable
(0-44%).5,21-26 As in the present study, they are usually seen
more frequently in surgically resected specimens than in
colonoscopic biopsy material, reflecting the predominance
of granuloma in the deeper layers.26 Furthermore, they may
be seen only in regional lymph nodes and would therefore
not be included in an endoscopic biopsy specimens.28,29 In
the absence of caseation, Crohn’s disease has to be
considered, as clinically it closely mimics intestinal TB. The
granulomas associated with TB have histological features
that are distinct from those of Crohn’s disease. The former
are frequently large and multiple, tend to coalesce, often
contain Langhans’ giant cells, may caseate, and contain AFB;
whereas in Crohn’s disease they are often small, infrequent
and never coalesce or caseate.30 Disappointingly, as in
the case of caseating granulomas the identification of
AFB in colonoscopic biopsy specimens has likewise been
reported with variable frequency (0-36%).6,22-26 Sending
specimens for culture of mycobacteria increases the yield
but takes time.21,23,31 More recently, polymerase chain
reaction assays have been applied to endoscopic biopsy
specimens, to facilitate rapid detection of mycobacterial
The management of GI TB relies on anti-tuberculous
drugs, surgery being reserved for complications or
uncertainty in diagnosis. Ideally, the diagnosis should be
confirmed by histology or isolation of mycobacteria.
However, the typical granulomas and AFB are not
invariably detected in the affected tissues, and a trial of
anti-tuberculous therapy may be indicated, whenever
clinical suspicion of the infection is high. A 6 to 9-month
course of anti-tuberculous chemotherapy is sufficient for
immunocompetent patients treated with a standard regimen
consisting of four first-line drugs. Typically this involves:
isoniazid, rifampicin, ethambutol, and pyrazinamide for the
initial 2 months, followed by isoniazid and rifampicin for
another 4 to 7 months.33,34 A longer duration of therapy is
necessary if one or more of these four first-line drugs
cannot be used because of intolerance or drug resistance.33
Surgical therapy is usually reserved for intestinal TB
complicated by perforation, obstruction, or uncontrollable
haemorrhage.1,2,8-10 If surgery is to be performed, the most
conservative approach should be used.1,2,9,10 Tuberculous
perforations and obstructing lesions warrant surgical
resection.18 Nevertheless, conservative treatment has been
advocated for tuberculous intestinal obstruction in stable
patients, as most will improve with anti-tuberculous
chemotherapy; surgery being carried out for those who fail
to respond.14,35 Remarkably, two of our three cases of
tuberculous intestinal obstruction resolved with conserva-
tive treatment, even prior to the diagnosis of TB (by
colonoscopy) and the institution of anti-tuberculous
An increased number of patients with intestinal TB have
been encountered in recent years. Due to the lack of
characteristic specific symptoms and signs and the fact
that laboratory findings are commonly non-specific or
show no abnormality, diagnosing such cases is difficult. A
heightened awareness of this potentially fatal yet readily
treatable condition is therefore important. A notable clue
is the presence of abdominal pain co-existing with active
pulmonary TB, but such lung involvement is present in less
than one fifth of the cases. The most common site of infec-
tion is the ileocaecal region, for which colonoscopy with
ileoscopy plus biopsy is the diagnostic tool of choice.
Surgery is usually reserved for patients with complications
or those whose diagnosis cannot be ascertained by other
1. Marshall JB. Tuberculosis of the gastrointestinal tract and peritoneum.
Am J Gastroenterol 1993;88:989-99.
Horvath KD, Whelan RL. Intestinal tuberculosis: return of an old
disease. Am J Gastroenterol 1998;93:692-6.
Corbett EL, Watt CJ, Walker N, et al. The growing burden of
tuberculosis: global trends and interactions with the HIV epidemic.
Arch Intern Med 2003;163:1009-21.
Tuberculosis notifications and rate by age and sex 2005.
Tuberculosis in Hong Kong website: http://www.info.gov.hk/dh/
publicat/web/tb/tb2005e.htm. Accessed 25 Feb 2006.
Tuberculosis and Chest Service of the Department of Health
Annual Report 2004. Tuberculosis in Hong Kong website: http://www.
info.gov.hk/tb_chest/doc/Annualrept_2004.pdf. Accessed 25
Patel N, Amarapurkar D, Agal S, et al. Gastrointestinal luminal
tuberculosis: establishing the diagnosis. J Gastroenterol Hepatol
Global tuberculosis control: surveillance, planning, financing. Geneva:
World Health Organization; WHO report 2005.
Horsburgh CR, Nelson AM. Mycobacterial disease of the
gastrointestinal tract. In: Blaser MJ, Smith PD, Ravdin JI, Greenberg
HB, Guerrant RL, editors. Infections of the gastrointestinal
tract. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2002:
Wang HS, Chen WS, Su WJ, Lin JK, Lin TC, Jiang JK. The
changing pattern of intestinal tuberculosis: 30 years’ experience.
Int J Tuberc Lung Dis 1998;2:569-74.
10. Faylona JM, Chung SC. Abdominal tuberculosis revisited. Ann
Coll Surg 1993;3:65-70.
11. Tuberculosis and Chest Service of the Department of Health
Annual Report (1995-2004). Tuberculosis in Hong Kong website:
Accessed 25 Feb 2006.
12. Sharma MP, Bhatia V. Abdominal tuberculosis. Indian J Med Res
13. Bhansali SK, Desai AN, Dhaboowala CB. Tuberculous perforation
of the small intestine. A clinical analysis of 19 cases. J Assoc
Physicians India 1968;16:351-5.
14. Bhansali SK. Abdominal tuberculosis. Experiences with 300 cases.
Am J Gastroenterol 1977;67:324-37.
Hong Kong Med J Vol 12 No 4 August 2006 271 Download full-text
Intestinal tuberculosis in Hong Kong
15. Schulze K, Warner HA, Murray D. Intestinal tuberculosis: experience
at a Canadian teaching institution. Am J Med 1977;63:735-45.
16. Seabra J, Coelho H, Barros H, Alves JO, Goncalves V, Rocha-Marques
A. Acute tuberculous perforation of the small bowel during
antituberculosis therapy. J Clin Gastroenterol 1993;16:320-2.
17. Scriven JM, Berry D. Multiple small bowel perforations in a patient
on treatment of tuberculosis. J R Coll Surg Edinb 1996;41:353.
18. Ara C, Sogutlu G, Yildiz R, et al. Spontaneous small bowel
perforations due to intestinal tuberculosis should not be repaired by
simple closure. J Gastrointest Surg 2005;9:514-7.
19. Leung VK, Chu W, Lee VH, Chau TN, Law ST, Lam SH. Tubercu-
lous intestinal perforation during anti-tuberculous treatment. Hong
Kong Med J 2006;12:313-5.
20. Hoon JR, Dockerty MB, Pemberton Jde J. Ileocecal tuberculosis
including a comparison of this disease with nonspecific regional
enterocolitis and noncaseous tuberculated enterocolitis. Int Abstr
21. Bhargava DK, Tandon HD, Chawla TC, Shriniwas, Tandon BN, Kapur
BM. Diagnosis of ileocecal and colonic tuberculosis by colonoscopy.
Gastrointest Endosc 1985;31:68-70.
22. Shah S, Thomas V, Mathan M, et al. Colonoscopic study of 50
patients with colonic tuberculosis. Gut 1992;33:347-51.
23. Bhargava DK, Kushwaha AK, Dasarathy S, Shriniwas, Chopra P.
Endoscopic diagnosis of segmental colonic tuberculosis. Gastrointest
24. Singh V, Kumar P, Kamal J, Prakash V, Vaiphei K, Singh K.
Clinicocolonoscopic profile of colonic tuberculosis. Am J
25. Kim KM, Kee A, Choi KY, Lee KY, Kwak JJ. Intestinal tuberculosis:
clinicopathologic analysis and diagnosis by endoscopic biopsy. Am J
26. Misra SP, Misra V, Dwivedi M, Gupta SC. Colonic tuberculosis:
clinical features, endoscopic appearance and management. J
Gastroenterol Hepatol 1999;14:723-9.
27. Leung VK, Tang WL, Cheung CH, Lai MS. Importance of ileoscopy
during colonoscopy for early diagnosis of ileal tuberculosis: report
of two cases. Gastrointest Endosc 2001;53:813-5.
28. Howell JS, Knapton PJ. Ileo-caecal tuberculosis. Gut 1964;19:524-9.
29. Tandon HD, Prakash A. Pathology of intestinal tuberculosis and its
distinction from Crohn’s disease. Gut 1972;13:260-9.
30. Pulimood AB, Ramakrishna BS, Kurian G, et al. Endoscopic
mucosal biopsies are useful in distinguishing granulomatous
colitis due to Crohn’s disease from tuberculosis. Gut 1999;45:537-41.
31. Bhargava DK, Shriniwas, Chawla TC, Tandon BN, Kapur BM.
Intestinal tuberculosis: bacteriological study of tissue obtained
by colonoscopy and during surgery. J Trop Med Hyg 1985;88:
32. Anand BS, Schneider FE, El-Zaatari FA, Shawar RM, Clarridge JE,
Graham DY. Diagnosis of intestinal tuberculosis by polymerase
chain reaction on endoscopic biopsy specimens. Am J Gastroenterol
33. Blumberg HM, Burman WJ, Chaisson RE, et al. American Thoracic
Society/Centers for Disease Control and Prevention/Infectious
Diseases Society of America: treatment of tuberculosis. Am J Respir
Crit Care Med 2003;167:603-62.
34. Kim SG, Kim JS, Jung HC, Song IS. Is a 9-month treatment sufficient
in tuberculous enterocolitis? A prospective, randomized, single-
centre study. Aliment Pharmacol Ther 2003;18:85-91.
35. Anand BS, Nanda R, Sachdev GK. Response of tuberculous stricture
to antituberculous treatment. Gut 1988;29:62-9.