ABSTRACT
Objective: To determine whether Mycobacterium tuberculosis infection spreads through the blood
to different lymph-node groups in patients with tuberculous lymphadenitis.
Design: Prospective analytical study.
Setting: The patients were recruited, managed and followed at the lymphodenopathy clinic, Central
Police Hospital, Burr, Khartoum, Sudan.
Subjects: Fifty two sequential patients were enrolled. Thirty patients with FNAC diagnosis
of tuberculous lymphadenitis and positive PCR for M. tuberculosis complex had a mean age
of 26.9±11.2 years and similar male, female affection. Nine patients with FNAC tuberculous
lymphadenitis, but negative PCR had a slightly higher mean age (32.6±18.2 years) with similar
male : female proportions. Patients with reactive lymphadenopathy (9/52) were older than patients
with tuberculous lymphadenitis with a mean age of 45±24.6 years.
Results: None of the patients were positive for HIV or had clinical or radiological evidence of
pulmonary tuberculosis. M.tuberculosis DNA was detected in the blood samples of 30/39 (77%)
patients with tuberculous lymphadenitis, but in none of the cases with reactive or malignant
lymphadenopathy. The presence of M.tuberculosis DNA correlated strongly to multiple lymphnode
involvement [OR (odds ratio) = 96.7, 95% confidence interval (CI) 9.0 – 1,039] and to caseatinggranulomatous
and predominantly necrotic cytomorphological categories [OR = 70, 95% confidence
interval (CI) 7.0 – 703].
Conclusion: M.tuberculosis most probably disseminates through the blood from one node group
to the other in patients with tuberculous lymphadenitis.
East African Medical Journal Vol. 84 No. 1 January 2007
HAEMATOGENOUS DISSEMINATION OF TUBERCULOUS LYMPHADENITIS
G.S. Sharafeldin, MSc, National Health Laboratory, Federal Ministry of Health, Khartoum, Sudan, E.A.G. Khalil, MBBS,
FRCPath, Institute of Endemic Diseases, University of Khartoum, P.O. Box 45235, Khartoum, Sudan, I.A. El Hag,
MBBS, PhD, PARAS Central Hospital, Sakaka Al-jouf, Saudi Arabia, K.E. Elsiddig, MBBS, MCS, FRCS, Departments
of Surgery and Pathology, Faculty of Medicine, University of Khartoum, P.O. Box 45235, Khartoum, Sudan, M.E.M.O.
Elsafi, MBBS, PhD, MD, Central Police Hospital, Khartoum, Sudan, A.S.A Aijafari MSc, A.A. Elnour MSc, Institute of
Endemic Diseases, University of Khartoum, P.O. Box 45235, Khartoum, Sudan, A.M. Hussein, MBBS, MD, Central Police
Hospital, Khartoum, Sudan, I.M. Elkhidir, MBBS, PhD, Department of Microbiology, Faculty of Medicine, University of
Khartoum, P.O. Box 45235, Khartoum, Sudan and A.M. El-Hassan, DKSM, PhD, FRCPath, Institute of Endemic Diseases,
University of Khartoum, P.O. Box 45235, Khartoum, Sudan
Request for reprints to: Prof. E.A.G. Khalil, Department of Clinical Pathology and Immunology, Institute of Endemic
Diseases, University of Khartoum, P.O. Box 45235, Khartoum, Sudan

INTRODUCTION
Tuberculosis remains one of the major health
problems worldwide with 8.7 million new cases
every year and an estimated 1.7 million annual
deaths (1). Tuberculosis is a systemic disease
affecting almost all organs with two clinical forms;
pulmonary and extra-pulmonary. The incidence
of extra-pulmonary tuberculosis is dramatically
rising, largely because of HIV/AIDS pandemic (2).
Tuberculous lymphadenitis is the most common
type of extra-pulmonary tuberculosis (3-6). Lymph4
E A S T A F R I C A N M E D I C A L J O U R N A L January 2007
nodes are usually involved as a component of
a primary complex. Less often lymphadenitis is
seen in secondary tuberculosis but the nodes are
usually smaller and firmer than in primary disease.
Tuberculous lymphadenitis may involve a single
discrete lymph-node, multiple lymph-nodes or
multiple sites of lymph-nodes and could even be
generalised. Nevertheless, the most commonly
affected lymph-nodes are those of the cervical
region (7,8). Since tuberculous lymphadenitis is
predominantly a primary disease, the question of how
the infection spreads from the primary lymph-node
to other nodes arises. Using blood culture techniques,
circulating M. tuberculosis has been found to be one
of the frequent causes of bloodstream infections
(BSI) among febrile adults with advanced HIV
infection in sub-Saharan Africa. These patients have
radiological, microbiological or clinical evidence for
active pulmonary disease. Recently, non-tuberculous
mycobacteria was detected by culture technique in
immunocompetent individuals (9-13).
In this communication we provide evidence
to the probable route of M. tuberculosis spread
in peripheral tuberculous lymphadenitis by
demonstrating circulating M. tuberculosis DNA in
immune-competent Sudanese patients who had no
signs of pulmonary disease, using a highly sensitive
molecular technique (PCR).

MATERIALS AND METHODS
The study proposal was scientifically and ethically
reviewed by the Ethical Committee of the Institute
of Endemic Diseases, University of Khartoum. Fifty
two sequential patients with lymphadenopathy
were enrolled in this prospective study following
informed consent. FNAC was performed on
all patients as a routine diagnostic procedure.
Following thorough clinical examination; blood
cell count, ESR, Mantoux test, HIV ELISA test
and a chest X-ray were performed. Five milliliters
of EDTA-blood were collected from all patients
and DNA was extracted from the mononuclear
blood cells (PBMCs) using the phenol-chloroform
iso-amyl alcohol (PCI) method. Peripheral blood
mononuclear cells (PMBCs) were collected using
density gradient centrifugation with FicolHypaque.
The quality of the extracted DNA was checked by
1.2% agarose gel electrophoress.
Oligonucleotide primers: A single set of oligonucleotide
primers was used (MOL BIOL, Berlin, Germany®).
The lyophilised primers were reconstituted as
described by the manufacturer. The target for PCR
amplification was IS6110 (size 123bp), an insertionlike
element found in Mycobacterium tuberculosis
complex strains. The oligonucleotides primers used
were:
E1 (20 mer)(5’-CCTGCGAGCGTAGGCGTCGG-3’)
E2 (5’ -CTCGTCCAGCGCCGCTTCGG-3’)
DNA amplification: Following strict Standard
Operating Procedures, DNA extraction and
PCR were carried in separate rooms to reduce
cross contamination. PCR for the specimens was
performed in a total volume of 50μl of the reaction
mixture containing 10X PCR buffer, 2.25μM.
Magnesium Chloride, dNTPs mixture (dATP, dGTP,
dCTP & dTTP) 100 each (Boehringer Mannheim,
Germany), E1 and E2 oligonucleotide primers 0.2μM
each and 2 U/μL of Taq polymerase (Finnzmes®,
Vienna, Austria). The reaction mixture was overlaid
with mineral oil. The tubes were then subjected to
40 thermal cycles in a programmable heat block
(Biometra®, Göttingen, Germany). The cycle was
as follows: denaturation at 95ºC for five minutes,
annealing at 65ºC for one minute, extension at 72ºC
for one minute and final extension at 72ºC for ten
minutes.
Detection of amplified DNA: The PCR products were
analysed by ethidium bromide-stained agarose
gel electrophoresis on 1.5% agarose gel. Positive,
negative controls and a 100 bp DNA marker were
included with every electrophoresis run.
Statistical analysis: Odds ratio were computed on
AcaStat statistical software. If any of the four values
in the contingency table were zero, one is added to
all values before calculating the odds ratio.

RESULTS
Baseline characteristics, laboratory findings, FNAC
and PCR results were summarised in Table l.
Cervical lymph nodes were the most commonly
affected group and were seen in 91% of patients
with lymphadenopathy. The cytological diagnosis
was; tuberculous lymphadenitis in 39 cases (75%),
reactive in nine (17.3%) and malignant in four cases
January 2007 E A S T A F R I C A N M E D I C A L J O U R N A L 5
(7.7%). Based on the cytological findings, cases of
tuberculous lymphadenitis were categorised into
three groups; necrotising-granulomatous (21/39;
54%), predominantly necrotic (6/39; 15.3%) and
granulomatous (12/39; 30.7%). Multiple lymphnodes
were seen in 28 cases, but a single discrete node
was seen in 11 cases of tuberculous lymphadenitis. All
cases in the caseating-granulomatous (21/21) and the
predominantly necrotic (6/6) categories had multiple
lymphnodes. Single lymph-node presentation was
mainly associated with granulomatous lesion
(11/12). An ear, nose, throat examination revealed
no abnormalities.
None of the patients were HIV positive or had
radiological or clinical evidence of pulmonary or
miliary disease.
The ESR was high in the tuberculous group
especially the granulomatous type, while it was
variable in patients with reactive and malignant
groups. A significant Mantoux reactivity (> l 5
mm) was seen in all patients with tuberculous
lymphadenitis, while it was < 10 mm in the reactive
and the malignant groups.
Circulating M. tuberculosis DNA was detected
in 30 out of 39 cases (77%) with tuberculous lymph
nodes, but in none of the cases with reactive or
malignant nodes. The PCR positive rates varied
among tuberculous patients with different smear
finding, while it was 100% (27/27 cases) in the
caseating-granulomatous and the predominantly
necrotic categories; the positivity rate was 25%
(3/12 cases) among patients with granulomatous
lesions. The computed odds ratio was 70 with a
95% confidence interval CI between 7.0 and 703.
Circulating M. tuberculosis DNA was detected in
all of the 28 patients (100%) with multiple lymph
nodes, but in 2 out of 11 cases (18%) with single node
involvement. The odds ratio was 96.7 with a 95%
confidence interval CI between 9.0 and l,039.

DISCUSSION
Tuberculous lymphadenitis is characterised
by painless enlargement of lymph nodes and
occasional constitution symptoms like nocturnal
fever and sweating. The cervical group is the most
commonly affected group. Peripheral tuberculous
lymphadenitis usually develops as a part of a primary
complex i.e. a primary focus and regional glands.
Involvement of lymph-nodes as a manifestation of
a generalised tuberculous infection is rather rare
(14). For tuberculous lymphadenitis of the neck,
the primary focus would be mainly in the tonsils
and mouth, however recent or previous seeding
of the lymph nodes from an occult site cannot be
ruled out. None of our cases had shown clinical or
radiological evidence of another primary complex
or military disease and the lymph-nodes were
large and soft. These findings probably support the
primary nature of the disease in our cases. Therefore,
the demonstration of circulating mycobacteria in
such cases would suggest spread by blood stream.
Patients with multiple nodes were considered as
having bulky disease.

Table 1
Baseline characterisitics, cytomorphological patterns and circulating mycobacterial DNA of the study patients
Cytomorphological type
of lymphadenopathy M: F Mean age Mean ESR Mean TBCs Mean mantoux
mm mm
PCR positive (n = 30) 1:1.5 26.9 ± 11.2 87 ± 31 5.3 ± 1.6 22 ± 6
Necrotising Tb (n = 6) 1:5 23 ± 8.6 69 ± 24.9 6.8 ± 2.3 19.2 ± 5.5
Granulomatous Tb (n = 3) 1:2 43.7 ± 15.1 113 ± 25.1 4.6 ± 0.7 36.3 ± 7.6
Necro/Granul.Tb (n = 21) 3:4 25.1 ± 9.2 69.3 ± 32.4 4.6 ± 0.6 21.4 ± 6
PCR negative (n = 22) 2:1 34.6 ± 21.6 77 ± 30 5.9 ± 3.2 13 ± 10.8
TB lymphadenitis (n=9) 2:1 32 ± 18.2 90.4 ± 27.8 7.3 ± 4.5 19.3 ± 8.2
Reactive (n = 9) 2:1 45 ± 24.6 76 ± 35.1 5.3 ± 1.7 5.3 ± 9.2
Malignancy (n = 4) 1:0 26.8 ± 25.5 56.7 ± 23.1 4.3 ± 0.6 8.5 ± 6.6
Continuous variables are expressed as means ±SD
6 E A S T A F R I C A N M E D I C A L J O U R N A L January 2007
For the detection of circulating M. tuberculosis,
a highly sensitive molecular technique (PCR)
was used in this study. PCR has proven to be
more sensitive compared to the conventional
microbiological methods (LJ culture and ZN
staining for acid fast bacilli in smears) in the
identification of mycobacterium (15-19). However,
PCR is known to be associated with high false
positive results, with rates ranging from 3 – 20%,
mainly due to cross-contamination (20, 21). To
prevent cross-contamination we followed strictly
the manufacturer instructions, DNA extraction and
amplification were carried out in different rooms
and the sequences of the process were adequately
monitored. Positive and negative controls were
included as extra quality assurance measure with
every run. We have recently shown that, over 96%
of cases of tuberculous lymphadenitis in Sudan
are caused by M. tuberculosis (18). This made the
use of a single primers set that amplifies a 123
bp sequence common to all M. tuberculosis more
appropriate for this study. Circulating M. tuberculosis
DNA could be demonstrated in most patients
with tuberculous lymphadenitis in this study. The
highest positive rate was reported among cases with
caseating-granulomatous or predominantly necrotic
changes. The positive rate among patients with
granulomatous lesion was low. This agrees well with
previous studies demonstrating lower mycobacterial
load in granulomatous lesions which was reflected
in scantier mycobacteria in ZN smears and lower
positive rates in LJ culture (22,23). Although
circulating M. tuberculosis DNA was seen in patients
with multiple lymph-node involvement or showing
caseating-granuloma tous or predominantly necrotic
cytological pictures, the confidence intervals were
wide probably indicating weak association. The
granulomatous pattern is associated with single
node involvement (91.6%) and a low positive rate
(25%) for circulating M. tuberculosis DNA. Many
studies showed that bloodstream infections (BSI) by
M. tuberculosis do occur. However, it affects mainly
immuno-compromised adults with advanced HIV
infection with radiological, microbiological and/
or clinical evidence of pulmonary tuberculosis.
The demonstration of circulating M. tuberculosis
DNA in immuno-competent patients, mainly in
association with multiple lymph node involvement
probably suggests haematogenous spread. Lack of
multiple organ involvement by the disease in-spite
of haematogenous spread can be explained by
compartmentalisation of mycobacterial infection.
Garcia de Viedma et al (24), demonstrated that
infection by more than one mycobacterial strains is
very rare and when occurs, the co-infecting strains
are not equally distributed at pulmonary and
extra-pulmonary sites. Circulating M. tuberculosis
DNA in our patients could explain the nocturnal
fever that was reported by more than 80% of our
patients (unpublished data). The pattern of the
fever could be explained by episodic release of the
mycobacteria and its antigens into the blood stream.
Demonstration of circulating M. tuberculosis DNA
was previously shown by Mirza et al (25).
It has long being claimed that the mantoux
test can help to differentiate between tuberculous
and non-tuberculous lymphadenitis (26,27). All
our patients with tuberculous lymphadenitis
had strongly positive mantoux test (induration
>15mm).
We conclude that the presence of circulating
M. tuberculosis DNA in the blood of most patients
with tuberculous lymphadenitis, especially those
with multiple lymph node involvement indicates
that heamatogenous spread is the probable route
of mycobacterial dissemination in tuberculous
lymphadenitis.

REFERENCES
1. World Health Organisation. (2002). Global Tuberculosis
Control: Surveillance, Planning, Financing. WHO report,
Geneva, Switzerland, 2000 WHO/CDC/TB/; 295.
2. Fanning A. Tuberculosis: Extrapulmonary disease. CMAJ.
1999; 160: 1597-1603.
3. Jha B.C., Dass A., Nagarkar N.M., Gupta R. and Singhal S.
Cervical tuberculous lymphadenopathy: changing clinical
pattern and concepts in management. Postgrad. Med. J. 2001;
77: 185-187.
4. Chao S.S., Loh K.S., Tan K.K. and Chong S.M. Tuberculous
and nontuberculous cervical lymphadenitis: a clinical review.
Otolaryngol. Head Neck Surg. 2002; 126: 176-179.
5. Shenoy R., Kapadi S.N., Pai K.P., et al. Fine needle aspiration
diagnosis in HIV-related lymphadenopathy in Mangalore,
India. Acta. Cytol. 2002; 46: 35-39.
6. El_Hassan A.M. and Tag El Deen M. Primary intestinal
tuberculosis in the Sudan. Ann. Trop. Med. Parasitol. 1982; 76:
317-322.
7. Bandapat M.C., Mishra B.M., Dash S.P. and Kar P.K.
Peripheral lymph-node tuberculosis: A review of 80 cases.
Brit. J. Surg. 1990; 77: 911-912.
January 2007 E A S T A F R I C A N M E D I C A L J O U R N A L 7
8. Yassin M.A., Olobo J.O., Kidane D., et al. Diagnosis of
tuberculous lymphadenitis in Butjira, Rural Ethiopia. Scand.
J. Infect. Dis. 2003; 35: 240-243.
9. Archibald L.K., den Dulk M.O., Pallangayo K.J. and Peller
L.B. Fatal Mycobacterium tuberculosis bloodstream infections
in febrile hospitalised adults in Dar es Salaam, Tanzania. Clin.
Infect. Dis. 1998; 26: 290-296.
10. Archibald L.K., McDonald L.C., Nwanyanwu O., et al. A
hospital-based prevalence survey of bloodstream infections
in febrile patients in Malawi: Implications of diagnosis and
therapy. J. Infect. Dis. 2000; 181: 1414-1420.
11. Waddell R.D., Lishimpi K., von Reyn C.F., et al. Bacteraemia
due to M. tuberculosis or Mbovis, Bacille Calmette Guerin
(BCG) among HIV-positive children and adults in Zambia.
AIDS. 2001; 15: 55-60.
12. Lai C.C., Lee L.N., Ding L.W., Yu C.J., Hsueh P.R. and
Yang P.C. Emergence of disseminated infections due to
nontuberculous mycobacteria in non-HIV-infected patients,
including immunocompetent and immunocompromised
patients in a university hospital in Taiwan. J. Infect. 2005;
[Epub ahead of print].
13. MacGregor R.R., Hafner R., Wu J.W., et al. ACTG Protocol
341 Team. Clinical, microbiological, and immunological
characteristics in HIV-infected subjects at risk for disseminated
Mycobacterium avium complex disease: an AACTG study.
AIDS Res. Hum. Retroviruses. 2005; 21: 689-695.
14. Miller F.J.W. and Cashman J.M. The natural history of
peripheral tuberculous lymphadenitis associated with a
variable primary focus. Lancet. 1955; 268: 1286-1289.
15. Singh K.K., Muralidhar M., Kumar A., et al. Comparison
of in house polymerase chain reaction with conventional
techniques for the detection of Mycobacterium tuberculosis
DNA in granulomatous lymphadenopathy. J. Clin. Pathol.
2000; 53: 355-361.
16. Goel M.M., Ranjan V., Dhole T.N., et al. Polymerase chain
reaction vs. conventional diagnosis in fine needle aspiration
of tuberculous lymph nodes. Acta. Cytol. 2001; 45: 333-340.
17. Hirunwiwatkul P., Tumwasorn S., Chantranuwat C. and
Sirichai U. A comparative study of diagnostic tests for
tuberculous lymphadenitis: polymerase chain reaction vs
histopathology and clinical diagnosis. J. Med. Assoc. Thai.
2002; 85: 320-326.
18. Aljafari A.S., Khalil E.A., Elsiddig K.E.G., et al. Diagnosis
of tuberculous lymphadenitis by FNAC, microbiologial
methods and PCR: A comparative study. Cytopathol. 2004;
15: 44-48.
19. Jain A., Verma R.K., Tiwari V. and Goel M.M. Dot-ELISA vs.
PCR of fine needle aspirates of tuberculous lymphadenitis: a
prospective study in India. Acta. Cytol. 2005; 49: 17-21.
20. Noordhoek G.T., Kolk A.H., Bjune G., et al. Sensitivity
and specificity of PCR for detection of mycobacterium
tuberculosis: A blind comparison study among seven
laboratories. J. Clin. Microbiol. 1994; 32: 277-284.
21. Noordhoek G.T., Mulder S., Wallace P. and Van Loon
A.M. Multicentre quality control study for the detection of
mycobacterium tuberculosis in clinical samples by nucleic
amplification methods. Clin. Microbiol. Infect. 2004; 10: 295-
301.
22. Gupta S.K., Chagh T.D., S.heikh Z.A. and
al-Rubah N.A. Cytodiagnosis of tuberculous lymphadenitis.
A correlative study with microbiological examination. Acta.
Cytol. 1993; 37: 329-332.
23. Prasson D. Acid fast bacilli in fine needle aspiration smears
from tuberculous lymph-nodes. Where to look for them. Acta.
Cytol. 2000; 44: 297-300.
24. Garcia de Viedma D., Martin M., Ruiz Serrano M.J., Alcala L.
and Bouza E. Polyclonal and compartmentalised infection by
Mycobacterium tuberculosis in patients with both respiratory
and extra-respiratory involvement. J. Infect. Dis. 2003; 187:
695-699.
25. Mirza S., Restrepo B.I., McCormick J.B. and Fisher-Hoch S.P.
Diagnosis of tuberculous lymphadenitis using a polymerase
chain reaction on peripheral blood mononuclear cells. Amer.
J. Trop. Med. Hyg. 2003; 69: 461-465.
26. Margileth A.M. The use of purified protein derivative
mycobacterial skin test antigens in children and adolescents:
purified protein derivative skin test results correlated with
mycobacterial isolates. Pediat. Infect. Dis. 1983; 2: 225-231.
27. Margileth A.M., Chandra R. and Altman R.P. Chronic
lymphadenopathy due to mycobacterial infection. Clinical
features, diagnosis, histopathology and management. Amer.
J. Dis. Child. 1984; 138: 917-922.