THE CONTRIBUTION OF FINE NEEDLE ASPIRATION BIOPSY IN
THE DIAGNOSIS OF MYCOBACTERIAL LYMPHADENOPATHY
WITH PARTICULAR REFERENCE TO CHILDREN
Dissertation presented by
COLLEEN ANNE WRIGHT
For a PhD degree in Anatomical Pathology
at
Stellenbosch University
August 2009
Promoter: Prof R.M.Warren
Co-promoter: Prof B.J.Marais
DECLARATION
I, the undersigned, hereby declare that the work contained in this dissertation is my own
original work and that I have not previously in its entirety or in part submitted it at any other
university for a degree.
Signature:
Date: 5 August 2009
In all first author papers the candidate obtained ethics consent, assisted with the sample
collection, collected the data, drew up the analysis plan and wrote the first draft of the
manuscript.
In addition the candidate was involved in all the microscopic evaluations and learnt the
laboratory techniques used in the molecular analyses.
In all the other papers the candidate helped conceptualise the study, assisted with sample
collection and evaluation of microscopic slides and made a substantial contribution to the
manuscript.
i
SUMMARY
Expediting a diagnosis of tuberculosis in children, particularly those who are
immunocompromised due to HIV/AIDS, is essential, as they are vulnerable to develop severe
forms of disease due to their immature or compromised immune systems. A significant
percentage of children (8 to 10%) with TB have TB lymphadenitis, in isolation, or in
combination with other disease manifestations.
Fine needle aspiration biopsy (FNAB) is a simple and minimally invasive procedure
well tolerated by children. It may be performed as an outpatient procedure by clinicians as
well as nurses, and excellent results can be achieved with training in the correct procedure.
The aim of this dissertation was to demonstrate that FNAB may contribute
significantly to the diagnosis of mycobacterial lymphadenitis, with particular reference to
children TB suspects. We first established that TB lymphadenitis is a common clinical
problem in children in TB endemic areas and that FNAB is an efficient simple and effective
diagnostic modality in children with peripheral lymphadenopathy.
We then proceeded to document the diagnostic yield and time to diagnosis of FNAB
compared to conventional laboratory specimens collected in children.
We investigated the value of additional diagnostic modalities such as autofluorescence
in improving the ability of cytology to make a definitive diagnosis of mycobacterial infection
based on cytomorphology and identification of the organism.
In countries where organisms such as Mycobacterium bovis BCG and nontuberculous mycobacteria are prevalent, culture with subsequent speciation is essential. The
amount of material harvested during FNAB is minuscule, and requires immediate bedside
inoculation for optimal yields.
We developed an inexpensive and effective transport medium to facilitate
mycobacterial culture from FNAB, even if this is collected at an outside facility. It is ideally
ii
suited for use in clinics and rural hospitals as it is stable at room temperature, maintains
viability of the organism for seven days, and the closed lid format reduces contamination.
Mycobacterial culture even using liquid-based media, takes up to 6 weeks, and this
delay is unacceptable particularly in children. We developed a Nucleic Acid Amplification
Technique (NAAT) using High Resolution Melt Analysis and applied this novel technique to
FNAB specimens submitted in transport medium. Although sensitivity remained suboptimal,
the technique is highly specific, simple and rapid. Its use could be incorporated into routine
microbiology laboratories, to assist with rapid diagnosis while cultures are pending.
We collected a solid body of evidence, which will promote the use of FNAB in
suspected mycobacterial lymphadenopathy, particularly in children in resource-limited
countries. The utilisation of the diagnostic methods identified will expedite speciation and
allow early and appropriate initiation of therapy. This is in keeping with Millennium
Development Goal 6: to combat TB by early detection of new cases and effective treatment.
iii
OPSOMMING
Kinders met tuberkulose (TB), en veral diegene met gekompromiteerde immuniteit as gevolg
van MIV/VIGS, het ‘n verhoogde neiging om ernstige siektebeelde te ontwikkel vanweë hul
onvolwasse of gekompromiteerde immuunsisteme. ‘n Spoedige diagnose van TB in kinders is
dus noodsaaklik. ‘n Betekenisvolle persentasie van kinders (8 tot 10%) met TB het TB
limfadenitis met of sonder meegaande ander siekteverskynsels.
Fynnaaldaspirasiebiopsie (FNAB) is ‘n eenvoudige en minimale indringende
prosedure wat geredelik deur kinders aanvaar word. Geneeshere en verpleegkundiges wie
toepaslike opleiding in die uitvoering van FNAB ontvang het, kan die prosedure op
buitepasiënte uitvoer en uitstekende resultate behaal.
Die doel van hierdie studie was om aan te toon dat FNAB betekenisvol kan bydra tot
die diagnose van mikobakteriële limfadenitis in veral kinders met vermoedelike TB. Daar was
eerstens bevestig dat TB limfadenitis ‘n algemene kliniese probleem is in kinders in TB
endemiese areas en dat FNAB ‘n doeltreffende, eenvoudige en effektiewe diagnostiese
modaliteit is in kinders met perifere limfadenopatie.
Vervolgens was FNAB se diagnostiese opbrengs en die tydsverloop tot diagnose
vergelyk met dié van konvensionele laboratoriummonsters wat in kinders verkry word.
Die bydrae van verdere diagnostiese modaliteite soos outofluoressensie tot ‘n
verbetering in sitologie se rol in die diagnose van mikobakteriële infeksie, soos gebaseer op
sitomorfologie en identifisering van organismes, is ondersoek.
In lande waar organismes soos Mycobacterium bovis BCG en nie-tuberkuleuse
mikobakterië heersend is, is kultuur en spesiebepaling noodsaaklik. Die hoeveelheid materiaal
wat met FNAB verkry word is baie min en vereis onmiddellike okulasie vir die beste
resultate.
iv
Tydens hierdie studie is ‘n goedkoop en effektiewe vervoermedium ontwikkel om
mikobakteriële kultuur van FNAB verkreë monsters te fasiliteer, selfs al is die monster vanaf
‘n buite fasiliteit bekom. Die vervoermedium is baie geskik vir gebruik in klinieke en
plattelandse hospitale. Dit is stabiel by kamertemperatuur, handhaaf lewensvatbaarheid van
organismes vir sewe dae, en die geslote dekselformaat verminder kontaminasie.
Mikobakteriële kultuur neem tot ses weke, selfs met die gebruik van
vloeistofgebaseerde mediums. Sodanige vertraging in die diagnose is veral in kinders
onaanvaarbaar. Tydens hierdie studie is ‘n Nukleïnsuur Amplifikasietegniek ontwikkel deur
die aanwending van Hoë Resolusie Smeltanalise en is hierdie nuwe tegniek toegepas op
FNAB verkreë monsters wat in die vermelde vervoermedium versamel was. Alhoewel
sensitiwiteit nie optimaal was nie, is die tegniek baie spesifiek, eenvoudig en vinnig. Dit kan
in roetine mikrobiologie laboratoriums gebruik word om vinnige diagnose te bewerkstellig
terwyl daar gewag word vir die kultuur se resultaat.
Hierdie studie bied omvattende bewys ter ondersteuning van die gebruik van FNAB
in veral kinders met vermoedelike mikobakteriële limfadenopatie in lande met beperkte
hulpbronne. Die toepassing van die diagnostiese metodes wat in hierdie studie identifiseer is
sal spesiebepaling bespoedig en vroegtydige en toepaslike behandeling verseker. Dit stem
ooreen met Millennium Ontwikkelingsdoelwit 6: om TB te beveg deur vroeë opsporing van
nuwe gevalle en effektiewe behandeling.
v
ACKNOWLEDGEMENTS
This mortal world is fickle and unstable like unto a shifting shadow and the human life is like
unto a mirage and a reflection on the water. (Abdu'l-Baha)
Our lives, and what we achieve, are but a reflection of the lives around us, therefore
I would like to thank
All those who have assisted either directly or indirectly in the completion of these studies and
this project. If you are not mentioned by name in the list below forgive me, I will remain
indebted to you all. I hope the envisaged improved diagnosis and outcome for children with
tuberculosis will be a testament to this combined effort.
Prof Ben Marais and Prof Rob Warren, my supervisors, for their encouragement, insight and
guidance throughout the course of this dissertation. For giving so generously of their time and
often prioritizing my research needs above their own commitments.
Kim Hoek for the hours spent optimizing the molecular techniques in this study. I will always
be grateful to you for your assistance and your kindness.
Prof Gladwyn Leiman who started me on the long quest to establish Fine Needle Aspiration
Biopsy as an invaluable diagnostic tool, and whose legacy in gaining acceptance of this
modality in South Africa will live on forever.
Prof Paul van Helden, for allowing me access to expertise and infrastructure within his
department, and for his encouragement.
vi
Prof Johann Schneider, Prof Juanita Bezuidenhout and Dr Mercia Louw, who took over many
of my duties and freed me to complete this work.
Prof Peter Wranz, for his love for cytology, unfailing encouragement and support.
All the registrars in Anatomical Pathology, present and past, who uncomplainingly obtained
consent from patients and collected samples for the studies. I cannot sufficiently express my
appreciation for your participation.
All the staff in Anatomical Pathology who tolerated my absence during the periods that I
spent committed to completing the study.
Prof Martin Kidd for his statistical analysis, assistance and patience.
The National Health Laboratory Service, for providing me with the environment and
infrastructure to conduct this research. I sincerely hope the mechanisms and diagnostic
techniques that have evolved and which we envisage will be implemented in our laboratory
services contribute to patient care.
The Faculty of Health Sciences, for their support, particularly Prof Wynand van der Merwe
and Prof Usuf Chikte, who always believed I could do it. Thank you.
vii
FUNDING
This project was made possible by the following generous contributions:
The establishment fund Of the Dean of the Faculty of Health Sciences.
The National Health Laboratory Systems Trust fund.
Staff development bursary.
Division of Molecular Biology and Human Genetics, Stellenbosch University.
viii
DEDICATION
This work is dedicated to my family, without whose support and understanding
this would not have been completed.
Ariel, my son, who has survived a lifetime of a mother as a student and of whom
I am so proud.
Thandiwe, my daughter, beloved child of the light, who brings such joy into my
life.
Alethe, Richard and Jonathan, may your lives be blessed and protected
and
Steve, my husband, partner and friend.
You will always be the centre of my world.
And to the children of Africa who continue to bear the burden of this disease
May you say of us, that we tried…..
There can be no keener revelation of a society's soul than the way in which it
treats its children.
Nelson Mandela
ix
ABBREVIATIONS
AIDS
Acquired Immune Deficiency Syndrome
BCG
Bacillus Calmette-Guerin
CDC
Centers for Disease Control and Prevention
CI
Confidence interval
DOTS
Directly observed therapy, short-course
ELISA
Enzyme-linked immunosorbent assay
FNAB
Fine Needle Aspiration Biopsy
HAART
highly active anti-retroviral therapy
HIV
Human Immunodeficiency virus
HRM
High Resolution Melting
IRIS
Immune reconstitution syndrome
LED
Light emitting diode
LIP
Lymphocytic interstitial pneumonitis
MDR
Multi drug resistant
MGIT
Mycobacterial growth indicator tube
MODS
Microscopic observation drug susceptibility assay
M.tuberculosis
Mycobacterium tuberculosis
M.bovis-BCG
Mycobacterium bovis BCG
MVP
Mercury vapor lamp
NAAT
Nucleic acid amplification tests
NHLS
National Health Laboratory Services
NTM
Non - tuberculous mycobacteria
OR
Odds ratio
PBS
Phosphate buffered saline
PCR
Polymerase chain reaction
PPL
Persistent peripheral lymphadenopathy
SA
South Africa
TB (PTB)
Tuberculosis (Pulmonary tuberculosis)
TBH
Tygerberg Hospital
TST
Tuberculin skin test
WHO
World Health Organization
ZN
Ziehl-Neelson
XDR
Extensively drug resistant
x
TABLE OF CONTENTS
Chapter
Page
Chapter 1.
Introduction
1
Chapter 2.
Comprehensive literature review
7
Fine needle aspiration Biopsy - an undervalued diagnostic modality
8
in pediatric mycobacterium disease.
Chapter 3.
TB lymphadenitis a common clinical problem
Tuberculous Lymphadenitis as a Cause of Persistent Cervical
33
34
Lymphadenopathy in Children from a Tuberculosis-Endemic Area.
Chapter 4.
Diagnostic utility and feasibility of fine needle aspiration biopsy
39
Fine needle aspiration biopsy: diagnostic utility in resource-limited settings.
40
Fine Needle Aspiration Biopsy – a first line diagnostic procedure in
46
pediatric tuberculosis suspects with peripheral lymphadenopathy.
Chapter 5.
Optimizing cytopathological diagnosis
Auto fluorescence of Mycobacteria on Lymph Node Aspirates –
66
67
A glimmer in the dark.
Diagnosing Mycobacterial lymphadenitis in children using Fine Needle
71
Aspiration Biopsy: Cytomorphology, ZN staining and Autofluorescence –
making more of less.
Chapter 6.
Optimizing culture-based and molecular diagnosis
Mycobacterial Transport medium for routine culture of Fine Needle
78
79
Aspiration Biopsies
Combining Fine Needle Aspiration Biopsy and High Resolution Melt Analysis
to reduce diagnostic delay in Mycobacterial Lymphadenitis.
89
xi
Chapter 7.
Conclusion
110
1
CHAPTER 1
INTRODUCTION
Tuberculosis (TB) was declared a global emergency in 1993 by the World Health
Organisation (WHO) when it was estimated that a third of the world's population is infected
with TB.1 The developing world bears the brunt of this disease burden; 95% of TB cases and
98% of TB deaths occur in the developing world, with sub-Saharan Africa reporting the
highest TB incidence in the world.2 In 2005 there were an estimated 8.8 million new TB cases
globally and 7.2 million of these were in Asia and sub-Saharan Africa. Despite efforts to
contain the epidemic new TB cases increased by 6% between 2000 and 2005, predominantly
in sub-Saharan Africa and Asia.
The problem is compounded by the parallel human immunodeficiency virus (HIV)
pandemic with 33.2 million adults and children living with HIV in 2007, and 68% of these in
sub-Saharan Africa.3 Infection with the HIV virus is estimated to increase the lifetime risk of
a patient infected with M tuberculosis of developing TB by 10 fold to 50%.4 South Africa
with 5.5 million people leads the world in the number of patients living with HIV, as well as
TB/HIV co-infection.5
TB in childhood
TB is a major cause of childhood morbidity and mortality in developing countries.6, 7 It is
estimated that 11% of the 8.3 million new TB cases diagnosed globally in 2000 occurred in
children,8 and in endemic areas children contribute at least 15-20% of the total disease
burden.9 In the Western Cape Province, a high burden region in South Africa, the TB
notification rate in children less than 14 years of age was reported at 620 /100,000 in 2007.10
Official child TB figures almost certainly represent a gross underestimate due to
problems in obtaining an accurate diagnosis in resource-limited settings, poor record-keeping
and under reporting.8 Children rarely have sputum smear-positive TB, although this reduces
2
the risk of actively spreading the disease, it makes it more challenging to establish a definitive
diagnosis. Expectorated sputum specimens are difficult if not impossible to obtain in children
under the age of 7-8 years. Collection of induced sputum and gastric aspirate specimens are
possible, but difficult to collect and reported bacteriological yields are low (30-40%).8, 11
A clinical diagnosis of TB in children is problematical and although chest radiography
is regarded as the most practical test to provide a reasonably accurate diagnosis of
intrathoracic TB in children with suspicious symptoms, it has multiple limitations.12 In HIVinfected children the diagnosis of both intrathoracic and extra thoracic TB is difficult due to
underlying chronic pathology such as lymphocytic interstitial pneumonitis (LIP) or
bronchiectasis and opportunistic infections. Chest radiography can also not provide diagnostic
confirmation in children with extra thoracic TB. In a study in the Western Cape, South Africa
10.7% of children treated for active TB had extra thoracic disease manifestations only;
cervical TB lymphadenitis was the most common condition in these children being present in
48.6%. TB lymphadenitis is a common cause of persistent cervical adenopathy in TB endemic
areas. Persistent peripheral lymphadenopathy in HIV-infected children is a criterion for the
clinical classification of HIV itself. It may also be associated with multiple other pathological
entities, from infections to neoplasia.
Mycobacterium tuberculosis is not the only mycobacterial organism which involves
peripheral lymph nodes. In developed countries non-tuberculous mycobacteria (NT M) is the
commonest mycobacterial organism isolated and BCG vaccination may be associated with
local complications such as abscess formation at the vaccination site and regional
lymphadenitis. It is also the most common manifestation of immune reconstitution syndrome
(IRIS) if vaccinated HIV-infected infants are commenced on highly active anti-retroviral
therapy (HAART). It is therefore important to develop methods that will provide a rapid and
definitive mycobacterial diagnosis in children with persistent peripheral lymphadenopathy.
3
Fine needle aspiration biopsy (FNAB) is a specimen collection technique which is
widely used in the diagnosis of palpable masses including peripheral lymphadenopathy. Its
value in the diagnosis of mycobacterial lymphadenitis in adults is well documented. FNAB is
a simple effective and safe modality for obtaining a representative sample of material from a
lymph node and the diagnosis of mycobacterial adenitis can be confirmed utilising a number
of different investigations, including cytomorphology, specific stains to identify the organism,
culture and polymerase chain reaction (PCR).
FNAB is not widely used for the diagnosis of mycobacterial disease in children. As
this is a rapid and minimally invasive technique which can be performed on an outpatient
basis, in a primary health care setting it is ideal for use in resource-limited countries with the
highest incidence of mycobacterial disease as well as HIV infection.
Study objectives
The main study objective was to evaluate and develop laboratory-based diagnostic techniques
to establish a rapid, accurate and cost effective diagnosis of mycobacterial lymphadenitis in
children, using FNAB as the sampling method.We applied a stepwise approach using multiple
studies with specific aims to achieve this objective.
To establish that TB lymphadenitis is a common clinical problem in children from TB
endemic areas
This aim is addressed in chapter 3:
A prospective community-based study documented the contribution of TB
lymphadenitis to persistent lymphadenopathy among children who present to primary health
care clinics with persistent cervical adenopathy in a TB endemic area.
To evaluate the diagnostic utility and feasibility of FNAB in a resource limited setting
4
These aims are addressed in chapter 4:
A laboratory-based retrospective study documented FNAB adequacy rates among
various health care professionals who performed the procedure and described the spectrum of
pathology seen.
A laboratory-based retrospective study compared FNAB to more established specimen
collection.
A laboratory-based retrospective study evaluated the utility of FNAB in HIV-infected
children (supplementary manuscript).
To determine optimal methods for cytopathological diagnosis of mycobacterial disease
These aims are addressed in chapter 5 and 7:
A prospective laboratory-based study conducted in adults and children to determine
the value of incorporating autofluorescence into cytopathological evaluation of lymph node
aspirates in patients with suspected mycobacterial disease.
A prospective laboratory-based study in children assessing the ability of FNAB to
diagnose mycobacterial lymphadenitis in children, using cytomorphology, autofluorescence
on Papanicolaou stained smears, Ziehl-Nielsen (ZN) staining and/or culture.
A retrospective laboratory-based study to assess cytomorphological patterns of M.
bovis BCG and M. tuberculosis on FNAB (supplementary manuscript).
A retrospective laboratory-based study to evaluate fluorescence microscopy in the
diagnosis of mycobacterial infection using FNAB samples and a rewind able (power
independent) LED light source (supplementary abstract).
To explore novel methods for optimal culture-based and molecular diagnosis
These aims are addressed in chapters 6 and 7:
5
A prospective pilot study, collecting residual material from fine needle aspiration
biopsy in 200 µL of sterile water, to investigate the feasibility of utilising PCR amplification
techniques with fractionation in agarose gel and visualisation using ethidium bromide
directly from FNAB samples (supplementary study).
A prospective hospital-based study to assess the value of using a simple mycobacterial
transport medium for bedside inoculation of FNAB specimens.
A prospective hospital-based study utilizing FNAB combined with High Resolution
Melt Analysis to reduce diagnostic delay in patients mycobacterial adenitis.
A simplified technique using FNAB and collection of residual material on FTA paper
and real-time PCR/HRM for the diagnosis of mycobacterial infection (supplementary
manuscript in progress).
References
[1]
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;163:1009-1021.
[2]
World Health Organization. Global Tuberculosis Control: Surveillance, Planning,
Financing: WHO Report 2007. Geneva: World Health Organization; 2007. Report No.:
WHO/HTM/TB/2007.376.
[3]
Harries AD. HIV/AIDS: the long haul ahead [Editorial]. Int J Tuberc Lung Dis
2008;12:1347-1348.
[4]
Harries A, Maher D, Graham S. TB/HIV: A Clinical Manual. 2nd ed. Geneva: WHO
2004.
[5]
UNAIDS. Regional fact sheets. AIDS epidemic update - Regional Summary:Sub-
Saharan Africa. 2007 [cited 10 March 2009]; Available from:
http://data.unaids.org/pub/FactSheet/2008/epi07_fs_regionalsummary_subsafrica_en.pdf
6
[6]
Beyers N, Donald PR. A prospective evaluation of children under the age of five years
living in the same household as adults with recently diagnosed pulmonary tuberculosis. Int J
Tuberc Lung Dis 1997;1:38-43.
[7]
Marais BJ, Hesseling AC, Gie RP, Schaaf HS, Beyers N. The burden of childhood
tuberculosis and the accuracy of community-based surveillance data. Int J Tuberc Lung Dis
2006;10:259-263.
[8]
Marais BJ, Graham SM, Cotton MF, Beyers N. Diagnostic and management
challenges for childhood tuberculosis in the era of HIV. J Infect Dis 2007;196:76-85.
[9]
Marais BJ, Gie RP, Schaaf HS, et al. Childhood pulmonary tuberculosis: old wisdom
and new challenges. Am J Respir Crit Care Med 2006;173:1078.
[10]
Hesseling AC, Cotton MF, Jennings T, et al. High Incidence of Tuberculosis among
HIV-Infected Infants: Evidence from a South African Population-Based Study Highlights the
Need for Improved Tuberculosis Control Strategies. Clin Infect Dis 2009;48:108-114.
[11]
Zar HJ, Hanslo D, Appoles P. Induced sputum versus gastric lavage for
microbiological confirmation of pulmonary tuberculosis in infants and young children: a
prospective study.Lancet 2005;365:130-134.
[12]
Marais BJ. Tuberculosis in children. Pediatric Pulmonology 2008;43.322-329
7
CHAPTER 2
COMPREHENSIVE LITERATURE REVIEW
Fine needle aspiration Biopsy (FNAB) - an undervalued diagnostic modality in pediatric
mycobacterial disease
Colleen A Wright, Rob Warren, Ben J Marais
International Journal of Tuberculosis and Lung Disease – in press
8
Fine needle aspiration biopsy (FNAB) - an undervalued diagnostic modality in pediatric
mycobacterial disease
1
Colleen A Wright, 2 Rob Warren, 3Ben J Marais
1
Division of Anatomical Pathology, Department of Pathology, Stellenbosch University and
NHLS
2
NRF Centre of Excellence in Biomedical Tuberculosis Research / MRC Centre for
Molecular and Cellular Biology, Division of Molecular Biology and Human Genetics,
Stellenbosch University, South Africa
3
Department of Pediatrics and Child Health and Desmond Tutu TB Centre, Stellenbosch
University, South Africa
Unstructured summary
Mycobacterial disease, particularly tuberculosis, is an escalating problem in developing
countries, fuelled by the parallel HIV pandemic. In tuberculosis (TB) endemic countries
children carry a very high burden of disease, which may be unrecognised due to the difficulty
in making a diagnosis based on clinical, radiological or laboratory methods. One of the main
hurdles is the difficulty of obtaining adequate specimens for bacteriological confirmation of
disease in children.
TB lymphadenitis is the most common extra-pulmonary manifestation of TB and up to
up to 22% of children with persistent cervical lymphadenopathy and no local cause may have
TB adenitis Therefore fine needle aspiration biopsy (FNAB), which is a simple and safe
outpatient procedure that can be performed by nurses in resource-limited settings and
provides material for direct microscopy as well as culture and susceptibility testing, provides
an excellent opportunity to obtain bacteriologic confirmation. However, it remains a greatly
underutilized specimen collection modality.
This review provides a comprehensive overview of the diagnostic difficulties faced in the
diagnosis of paediatric TB in resource-limited settings and suggests ways to utilise FNAB
9
as a practical modality for the rapid and effective diagnosis of mycobacterial disease in the
significant subset of patients who present with peripheral lymphadenopathy. It also provides
detail on how best to perform the technique and suggests ways of making it more widely
available in resource-limited settings that carry the brunt of the paediatric TB disease burden.
The scale of the problem
Mycobacterium tuberculosis
In 2005, there were an estimated 8.8 million new tuberculosis (TB) cases globally, 7.4 million
in Asia and sub-Saharan Africa. More than 4000 people died every day from TB related
illnesses in that year, with the developing world bearing the brunt of this massive disease
burden; 95% of TB cases and 98% of TB deaths occurred in the developing world.1 What is
even more alarming is the projected scale of the epidemic to come; the World Health
Organization (WHO) has estimated that between 2000 and 2020 1 billion people will be
newly infected with M. tuberculosis, resulting in 200 million TB cases and 35 million deaths.2
TB is a major cause of childhood morbidity and mortality in developing countries.3-4
Accurate figures on the extent of paediatric TB are not available due to inadequate health
information systems in endemic countries and the limited attention paid to children who
contribute little to TB transmission within affected communities. WHO estimates of disease
incidence are based on sputum smear-positive cases, but more than 80% of children with TB
are sputum smear-negative. Indications are that children contribute at least 15-20% of the
total disease burden in endemic areas and that they suffer severe TB related morbidity and
mortality.5 Recognition of this high disease burden and the fact that highly effective
treatment is available has led to increased emphasis on the diagnosis and management of
paediatric TB. WHO research priority guidelines for paediatric TB identified the evaluation of
new techniques to improve the diagnosis and management of paediatric TB as an urgent
research priority.6
10
According to 2005 WHO figures, South Africa, despite its relatively good
infrastructure and health care services, reported a TB incidence of 600/100,000
population/year with a mortality rate of 71/100,000 population/year.1 The Western Cape
Province, a very high burden region in 2007 reported the adult TB incidence as 1005.7 per
100,000 compared to the National incidence rate of 739.6/100,000 population.7 In 2007 the
Western Cape Department of Health reported a TB notification rate of 620 per 100 000 in
children (0-14 years of age).8 A prospective observational study conducted in Cape Town
demonstrated that children less than 13 years of age contributed 13.7% of the total TB
caseload with a calculated TB incidence of 407/100,000/year; 52.2% of that in adults (≥ 13
years of age).
Mycobacterium bovis BCG
BCG vaccination is recommended by the WHO for infants in TB-endemic areas as it affords
protection against disseminated disease in children.9 However, BCG vaccination may be
associated with local complications, such as abscess formation at the vaccination site and/or
regional lymphadenitis. Prior to the emergence of HIV as a cause of immunosuppression,
disseminated disease was rare (0.19–2 cases/1 million vaccinated infants) and was associated
with congenital immunodeficiency syndromes.10
In South Africa, a change in vaccine policy in July 2000 from percutaneous Tokyo
strain BCG to intradermal Danish strain BCG vaccine has seen a number of adverse events
reported in both HIV-infected and immune competent children.9, 11-12 Distant or disseminated
disease has become a serious concern in human immunodeficiency virus (HIV)-infected
children. In 2007 it was estimated that 417 per 100 000 HIV-infected infants, are affected by
disseminated BCG disease,12 a rate that is about 1000 times the rate in HIV-uninfected
infants, and with a mortality rate of 75–86%.13
11
WHO recommended in 2007 that HIV-infected infants should not receive BCG
vaccination, 14 but this policy is very difficult to implement and may not be beneficial in
countries where both HIV and TB are endemic, as the HIV status of the baby is not known at
the time of routine vaccination (at birth). Where possible however every effort should be
made to introduce selective BCG vaccination based on known HIV status. 13
Non-tuberculous mycobacteria (NTM)
Non-tuberculous mycobacterial infection (NTM) is due to mycobacterial organisms other than
Mycobacterium leprae or members of the Mycobacterium tuberculosis complex 15 and differ
from the former in their habitat, contagiousness and susceptibility to chemotherapy.
In developed countries such as the USA, as the incidence of tuberculosis has fallen, NTM are
more frequently isolated than those of the M. tuberculosis complex. In immune competent
adults NTM infection presents as pulmonary disease in a patient with pre-existing chronic
pulmonary disease,15 whilst in immune competent children the organs involved are the lymph
nodes and the skin. These children are generally under the age of six years and are otherwise
healthy with no constitutional symptoms.16-17
In countries with a high prevalence of TB and HIV, localized or disseminated disease
from non-tuberculous mycobacteria may occur, but is uncommon.18 However, a case study
from Thailand showed that of the 153 HIV-infected children who had initiated antiretroviral
therapy, nine developed the immune reconstitution syndrome (IRIS) due to non-tuberculous
mycobacterial organisms.19 This highlights the need for bacteriological confirmation and
speciation in suspected cases of mycobacterial infection.
Contribution of HIV
One of the millennium development goals of the United Nations is to reduce TB-related
mortality to 50% of the 1990 levels. Increasingly it appears that this goal will not be achieved,
12
particularly in sub-Saharan Africa,20 due in part to HIV infection, which is a major contributor
to the escalating TB epidemic in this region and the rest of the world.21-22 It is estimated that
in 2007 about 420 000 new HIV infections occurred in children, predominantly in subSaharan Africa, through perinatal transmission, in a population vulnerable to early contact
with tuberculosis.23 A recent study in the Western Cape Province, South Africa, recorded a
TB incidence of 1596 (disseminated TB 240)/100 000 in HIV-infected and 66 (disseminated
14)/100 000 in HIV-uninfected infants.8
Autopsy studies in sub-Saharan Africa have shown that 30 to 40% of deaths in HIVinfected adults and up to 20% in HIV-infected children are due to TB.3 In Zambia, 3 Cote d’
Ivoire, 24 Malawi, 25 and South Africa, 22, 26 TB-related mortality is considerably higher in
HIV-infected compared to HIV-uninfected children. Despite this high disease risk
experienced by HIV-infected children, the majority of child TB cases are still in HIVuninfected children, as compared to adults relatively fewer children are HIV-infected and
young children are highly susceptible to develop TB irrespective of their HIV status.27 When
children are co-infected with HIV and TB however, they show more rapid progression of the
disease and increased morbidity and mortality.28 29
The diagnostic dilemma
Children rarely have sputum smear-positive TB. Although this reduces their risk of actively
spreading the disease, it makes it more challenging to establish a definitive diagnosis. In
contrast to adults where the sensitivity of sputum culture approximates 80-90%,young
children are unable to expectorate,alternative specimens such as gastric aspirates or induced
sputa are difficult to collect and culture yields are low (widely reported as 30-40%). 30-31
Chest radiography is regarded as the most practical test that provides a reasonably accurate
diagnosis of intra-thoracic TB in children with suspicious symptoms. 32 However, it has
numerous limitations including limited availability, expertise is required to interpret child
13
radiographs,33 and it has no value in the diagnosis of extra-thoracic TB. A recent survey
demonstrated that 10.7% of children treated for active TB had extra-thoracic disease
manifestations only.34 TB lymphadenitis was the most common form of extra-thoracic
tuberculosis in these children (50%).
Traditionally, three fasting gastric aspirates samples are collected on three consecutive
mornings, requiring hospitalization of the child and frequently the caregiver. In developing
countries this may create severe hardship as caregivers usually have more than one child in
their care. Caregivers may also compromise their employment by remaining in hospital for a
prolonged period of time. Hospitalization incurs cost and occupies beds which may be needed
for other children. Sputum induction using mobilization and hypertonic saline have shown
improved yield compared to gastric aspirates,30 but more recent studies demonstrated
equivalence with gastric aspirates.35 The technique requires specialized training and
equipment and may present a nosocomial transmission risk.27, 32, 36
Novel culture methods have been developed in an attempt to circumvent the slow
turnaround times, poor sensitivity and cost of conventional automated liquid broth systems.
The most feasible alternative to date has been the microscopic observation drug susceptibility
assay (MODS) that uses an inverted light microscope to rapidly detect “spindle and cord
formation” in selective broth culture that is indicative of mycobacterial growth.32 Phage
amplification assays that use bacteriophages to detect the presence of live M. tuberculosis
have been less successful. 27
There is abundant literature utilizing commercial and in-house PCR for the diagnosis
of mycobacterial infection in sputum and body fluids, such as cerebral spinal fluid, but results
in the literature are highly variable and have not been well validated in children. 27 Recent
reviews and meta-analyses of PCR in TB meningitis, pleuritis and sputum smear-negative
pulmonary TB demonstrated poor sensitivity.37 In patients with sputum smear-positive TB,
however, PCR may play an invaluable role in offering rapid species identification and
14
detection of drug resistance.32 The application in patients with sputum smear-negative or
extra-pulmonary TB requires further evaluation and use in low income countries remains
limited due to cost constraints..
T-cell assays that measure interferon- γ released by lymphocytes in peripheral blood
after exposure to M. tuberculosis specific antigens, have been hailed as reliable TB tests; two
commercial tests are available T-SPOT.TB and QuantiFERON –TB GOLD. However, study
results are highly variable and the experience in children remains limited. Current consensus
is that these tests are unable distinguish latent infection (a third of the world population is
latently infected) from active disease and adds little to the traditional tuberculin skin test
(TST). Tests are also too expensive and complex for routine use in low- income countries.27,
32, 36, 38
Fine Needle Aspiration Biopsy (FNAB)
History
FNAB has been used since the early 20th century to diagnose infectious and neoplastic
disease.39 Prior to the 1930s there were isolated case reports documenting its history, and the
first large scale report on needle aspiration biopsies was published from Memorial Hospital,
New York in 1930.40 In the 1960s this technique was championed by the Scandinavians and
was widely practiced, with many publications on its utility, mainly in the Scandinavian
literature.41 Gradually the use of FNAB spread through Europe and interest was rekindled in
the United States in the 1980’s. FNAB is now widely utilized as a first line diagnostic
procedure in the diagnosis of palpable masses, including peripheral lymphadenopathy. Its
value in the diagnosis of mycobacterial lymphadenitis in adults is well documented.42-44
FNAB is utilized throughout the developed and developing world, although the indications
differ in these two groups. In a review of MEDLINE by Das from 1966 to 2002, 39 849
journals published 5609 articles on FNAB, the number increasing sharply from the 1980s
15
onwards. Of the developed or high income countries, 52.8% (28/53) published articles
compared to 29.7% (46/155) of the low income or developing economies. South Africa was
one of the top 10 countries in the developing world with publications on FNAB. Although the
developed countries publish significantly more articles than the developing countries, these
tend to concentrate on breast and pancreatic disease, while the developing countries publish
significantly more articles on small round cell tumors and infectious diseases. This may
reflect the health care priorities and diagnostic needs of these developing countries.
Children with pulmonary TB have extra-thoracic disease manifestations in 10 to 30%
of cases. 45 2 TB lymphadenitis is the most common form of extra-thoracic TB in endemic
areas, 46 where up to 50% of extra-thoracic disease manifests as peripheral lymphadenopathy.
In fact, TB lymphadenitis is the most common (22-48%) cause of persistent cervical
lymphadenopathy in TB endemic areas.47-50 This is rarely recognized as a potential means for
specimen collection. Therefore it is important to develop methods that will provide a rapid
and definitive mycobacterial diagnosis in children with persistent cervical lymphadenopathy.
In these high burden TB endemic areas, where children may contribute up to 40% of the TB
caseload, 51 and assuming a conservative estimate of 5-10% of children with TB have
peripheral lymphadenopathy, FNAB could make a significant contribution to the diagnosis of
TB and other mycobacterial infections.
A recent retrospective study comparing FNAB to gastric aspirates and induced sputum in
children with pulmonary TB and peripheral lymphadenopathy showed FNAB to have a
superior diagnostic yield and a significantly reduced time to diagnosis. 52 An accurate
bacteriological diagnosis was made within 3 days in the vast majority of patients, which has
important benefits for patient management. FNAB should be regarded as a first line
diagnostic modality in paediatric TB suspects with peripheral lymphadenopathy.
TB and HIV have many features in common which contribute to diagnostic
difficulties, particularly in children. In this population both are chronic diseases that
16
commonly present with pulmonary symptoms/signs and/or lymphadenopathy.53 Persistent
peripheral lymphadenopathy is itself a criterion for the classification of HIV related disease in
children.54 Radiological features may be impossible to interpret with certainty and in children
with CD4 T-cell depletion, alternative tests such as the TST and or novel T-cell assays offer
little diagnostic assistance due to poor sensitivity.55
A study from South Africa demonstrated that the majority of HIV-infected children
with persistent lung disease have persistent peripheral lymphadenopathy (PPL) and PPL was
present in many children with pulmonary TB.56 In HIV-infected children, peripheral
lymphadenopathy may be associated with many pathological entities, from infections to
neoplasia. 56 Thus, enlarged peripheral lymph nodes were identified as important specimen
collection sites to consider, in establishing a definitive diagnosis of TB in HIV-infected
children. Another important disease presentation results from the immune reconstitution
inflammatory syndrome (IRIS), which may exacerbate the symptoms and signs of
mycobacterial disease and needs to be distinguished from TB treatment failure. 57 55 BCGrelated right sided axillary adenitis is the most common IRIS manifestation in BCGvaccinated infants recently started on highly active anti-retroviral therapy (HAART).
FNAB in children
FNAB has been less widely utilized in pediatrics as a diagnostic modality, although the
literature demonstrates a steady increase in the use of FNAB in children, achieving sensitivity
and specificity rates of over 90% in these studies. 58-60 The majority of the studies were in
pediatric oncology patients, where they play a role in triage, as childhood malignancy can be
difficult to diagnose, with signs and symptoms which often mimic other common paediatric
conditions especially viral infections. With significant mass lesions, FNAB proved to be the
diagnostic tool of choice in the triage of these patients.61 However, in developing countries
with a high burden of infectious diseases such as TB and HIV, FNAB can be of inestimable
17
value in confirming the diagnosis of mycobacterial infection, permitting early appropriate
therapy, as well as a means to obtain specimens for culture, bacterial species determination
and sensitivity testing. 60, 62-64
FNAB is a simple and minimally invasive technique, performed at the bedside, and is
well tolerated by children. Table 1 provides a summary of the equipment required and how to
perform the technique. Children under the age of six years are given are oral or intranasal
sedation for amnesia, but children older than six years tolerate aspiration with no sedation.
The technique is explained to the child, the caregiver is asked to remain with the child during
the procedure and the permission of the child is requested to proceed. If the child understands
that this is a very quick minimally painful procedure that will enable them to return home
after a short period of observation, they are willing to cooperate.
Most aspirates in children are from axillary or cervical lymph nodes that are easily accessible.
If the procedures are performed correctly, with a small gauge needle (22-23 G or smaller)
complications such as a small hematoma are rare. If the correct technique is utilized the yield
as well as the acceptance of the procedure by the patients will be excellent and nursing sisters
can be taught this technique with excellent results. In developing countries where TB is
endemic and specimens for bacteriological diagnosis are difficult if not impossible to obtain
from children, FNAB provides the means for accurate diagnosis in a significant percentage of
children with mycobacterial infection. This is particularly important in countries with a high
HIV rate, as well as increasing rates of multi drug-resistant (MDR) and extensively drugresistant (XDR) TB.
FNAB is a cost-effective diagnostic modality that may be performed at the bedside as
an outpatient procedure.62 It requires no infrastructure or sterile environment and no
sophisticated or expensive equipment. Superficial aspirates may be performed by clinicians,
pathologists or trained nursing personnel and do not require local or general anesthesia.61
18
FNAB is defined as an aspirate done using a cutting needle no larger than 22G.65 This
ensures minimal complications, the commonest of which is a small hematoma and thus FNAB
is the safest method of obtaining a tissue diagnosis.
A number of different techniques have been used to obtain a tissue and/or
bacteriological confirmation of TB lymphadenitis. These include core needle biopsy,66 wide
needle aspiration biopsy 66-67 and surgical lymph node biopsy.50, 56 None are as simple, safe
and cost effective to perform, as FNAB (Table 2). Compared to core needle biopsy, the risk
of sinus or fistula formation, permanent damage to nerves or seeding of tumor along the
needle track is minimal. 65 It permits sampling of the entire mass (node) as multiple
excursions of the needle are performed in the lesion without withdrawing the needle, and is no
more painful than venepuncture. No more than one ml of suction is applied during the
aspiration as the cutting needle will ensure material enters the needle by capillary action,
while more suction would increase the risk of bleeding and hematoma formation.
Most aspirates yield a very small amount of material which is expressed onto a glass
slide and a monolayer smear prepared. Preparation and staining of the smears can be done in a
very basic laboratory using manual methods. Although screening and reporting of the
aspirates should be performed by a cytopathologist. The smears are robust and may be
transported or even couriered to a central laboratory service for diagnosis and reporting.
FNAB is an effective and safe modality for obtaining a representative sample of material from
a lymph node, but laboratory confirmation of mycobacterial adenitis can be made utilising a
number of different investigations.
Laboratory analysis
Cytomorphology
Cytomorphology is simple, but not specific, as other opportunistic infections may present
with similar cytological changes and the quality of the smear is operator dependent. 42-43 A
19
number of reactions have been described in the literature: epithelioid granulomas without
necrosis, epithelioid granulomas with necrosis and necrosis without granuloma formation. 68
Ziehl-Neelson (ZN) staining
ZN staining, even in good laboratories, is insensitive, identifying organisms in only about
20% of culture positive samples. 44 Identification of the mycobacterial organism may be
dependent on the immune status of the patient and is reflected in the cytomorphological
pattern. Smears with poor granuloma formation are more likely to be positive on ZiehlNeelsen staining than those showing good epithelioid granulomas without necrosis. 68
Autofluorescence
Autofluorescence of mycobacteria is simple, sensitive and inexpensive, but is not widely
used. It is not highly specific, as the specificity is dependent on the experience of the operator
and it cannot differentiate between the various mycobacteria. 62, 69 The recent development of
light-emitting diode (LED) technology provides a cheap and reliable light source with a
usable life of more than 50,000 hrs using a fraction of the energy required by conventional
light bulbs and can run on batteries or inexpensive low-voltage power supplies. Preliminary
studies have indicated that LED-fluorescence microscopy offers a valid alternative to
conventional fluorescence microscopy 70 and studies are underway to investigate the
diagnostic utility of LED-auto fluorescence microscopy to detect Mycobacteria in fine needle
aspiration samples.
Culture
Direct inoculation into MGIT or Bactec tubes (Becton Dickinson, USA) at the bedside gives a
high yield, 62 but it takes 1-6 weeks for a result, and fails to differentiate the various
mycobacteria. Additional PCR-based testing is required for accurate speciation.. Direct
20
inoculation however is not feasible as the MGIT and Bactec collection media are expensive,
and are not readily available in clinics and wards for use by clinicians. We have developed an
inexpensive transport bottle which is easy to prepare “in house” and does not require
refrigeration.71 The positive culture yield and the time to positive culture from the transport
medium, even after 7 days at room temperature, is statistically no different to direct bedside
inoculation. The production and distribution of these bottles to clinics and hospitals, will
increase the feasibility of diagnosing tuberculosis at primary health care level
Nucleic acid amplification tests (NAATs)
A small number of recent studies have shown considerable promise in the use of NAATs in
the diagnosis of M. tuberculosis on FNAB specimens. 72-75 These studies have been
predominantly in adults. There is limited data on the use of FNAB for the diagnosis of
mycobacterial disease in children. Rapid and minimally invasive techniques such as FNAB,
which can be performed on an outpatient basis, in a primary health care setting, when
combined with a rapid and sensitive diagnostic technique such as PCR, may substantially
contribute to the effective management of mycobacterial infection in HIV-infected and
uninfected children. The implications of a rapid and accurate diagnosis include expediting
access to appropriate and adequate therapy as well as potentially limiting further costly
investigation.
21
Conclusion
It is difficult to understand why such an effective, simple and safe diagnostic modality is so
underutilized in the diagnostic workup of pediatric tuberculosis. It is ideally suited for
widespread use in resource limited countries which carry the dual burden of TB and HIV.
FNAB may be perceived as a procedure to be carried out by specialized medical personnel,
but it can safely be carried out by junior medical staff and nursing staff, capable of and trusted
with inserting intravenous lines. It has been demonstrated that samples collected by nurse
aspirators trained during a Burkitt’s lymphoma study conducted in Malawi is of similar
quality as those obtained by trained pathologists.
Adequate training is essential to develop confidence in the technique and to ensure a
good yield.56 This can be done through initial outreach teaching and training programs,
followed by cascade training to facilitate widespread dissemination of FNAB skills.
Pathology departments from South Africa and organizations such as the British Division of
the International Academy of Pathology and the Friends of Africa through the United States
and Canadian Academy of Pathology (USCAP) offer ongoing education in FNAB and
pathology in African and other TB-endemic countries. Most training programs incorporate
tuition in the cytopathological diagnosis of mycobacterial infection by pathologists, as the
advent of HIV has changed the cytomorphological appearance of these infections.
In addition to making a rapid and definitive tissue diagnosis, widespread use of FNAB
would provide ready access to material ideally suited to develop alternative methods for rapid
detection of mycobacteria, species determination and drug resistance testing. In order to
reduce the massive TB disease burden suffered by children in endemic areas we have to
utilize creative and effective means to ensure rapid and accurate diagnosis, such as FNAB,
that would facilitate early and appropriate treatment.
22
REFERENCES
[1]
World Health Organization. Global Tuberculosis Control: Surveillance, Planning,
Financing: WHO Report 2007. Geneva: World Health Organization; 2007. Report No.:
WHO/HTM/TB/2007.376.
[2]
Shingadia D, Novelli V. Diagnosis and treatment of tuberculosis in children. Lancet
Infect Dis 2003;3:624-632.
[3]
Chintu C, Mudenda V, Lucas SB. Lung diseases at necropsy in African children dying
from respiratory illnesses; a descriptive necropsy study. Lancet 2002;360:985-990.
[4]
McNally LM, Jeena PM, Gajee K, et al. Effect of age, polymicrobial disease, and
maternal HIV status on treatment response and cause of severe pneumonia in South African
children: a prospective descriptive study. The Lancet 2007;369:1440-1451.
[5]
Marais BJ, Gie RP, Schaaf HS, et al. Childhood pulmonary tuberculosis: old wisdom
and new challenges. Am J Respir Crit Care Med 2006;173:1078.
[6]
World Health Organization. A research agenda for childhood tuberculosis. Geneva:
World Health Organization; 2007. Report No.: WHO/HTM/TB/2007.381.
[7]
Health Systems Trust. 2007 [cited 6 March 2009]; Available from:
http://www.hst.org.za/healthstats/16/data
[8]
Hesseling AC, Cotton MF, Jennings T, et al. High Incidence of Tuberculosis among
HIV-Infected Infants: Evidence from a South African Population-Based Study Highlights the
Need for Improved Tuberculosis Control Strategies. Clin Infect Dis 2009;48:108-114.
[9]
Hesseling AC, Schaaf HS, Hanekom W, et al. Danish Bacille Calmette-Guerin
Vaccine–Induced Disease in Human Immunodeficiency Virus–Infected Children. Clin Infect
Dis 2003;37:1226-1233.
[10]
Casanova JL, Blanche S, Emile JF, et al. Idiopathic disseminated bacillus Calmette-
Guerin infection: a French national retrospective study. Pediatrics 1996;98:774-778.
23
[11]
Hesseling AC, Rabie H, Marais BJ, et al. Bacille Calmette-Guérin Vaccine–Induced
Disease in HIV-Infected and HIV-Uninfected Children. Clin Infect Dis 2006;42:548–558.
[12]
Hesseling AC, Marais BJ, Gie RP, et al. The risk of disseminated Bacille Calmette-
Guerin (BCG) disease in HIV-infected children. Vaccine 2007;25:14-18.
[13]
Mak TK, Hesseling AC, Hussey GD, Cotton MF. Making BCG vaccination
programmes safer in the HIV era. Lancet 2008;372:786-787.
[14]
Safety. GACoV. Revised BCG vaccination guidelines for infants at risk for HIV
infection; 2007 accessed 08 03 2009.
[15]
Piersimoni C, Scarparo C. Pulmonary infections associated with non-tuberculous
mycobacteria in immunocompetent patients. Lancet Inf Dis 2008;8:323-334.
[16]
Hazra R, Robson CD, Perez-Atayde AR, Husson RN. Lymphadenitis Due to
Nontuberculous Mycobacteria in Children: Presentation and Response to Therapy. Clin Infect
Dis 1999;28:123-129.
[17]
Flint D, Mahadevan M, Barber C, Grayson D, Small R. Cervical lymphadenitis due to
non-tuberculous mycobacteria: surgical treatment and review. Int J Pediatr Otorhinolaryngol
2000;53:187-194.
[18]
Zar HJ. Chronic lung disease in human immunodeficiency virus (HIV) infected
children. Pediatr Pulmonol 2008;43:1-10.
[19]
Puthanakit T, Oberdorfer P, Ukarapol N, et al. Immune reconstitution syndrome from
nontuberculous mycobacterial infection after initiation of antiretroviral therapy in children
with HIV infection. Pediatr Infect Dis J 2006;25:645.
[20]
Chaisson RE, Martinson NA. Tuberculosis in Africa - combating an HIV-driven crisis.
N Engl J Med 2008;358:1089-1092.
[21]
Braun MM, Cauthen G. Relationship of the human immunodeficiency virus epidemic
to paediatric tuberculosis and Bacillus Calmette-Guerin vaccination. Pediatr Infect Dis J
1992;11:220-227.
24
[22]
Madhi SA, Huebner RE, Doedens L, et al. HIV-1 coinfection in children hospitalised
with tuberculosis in South Africa. Int J Tuberc Lung Dis 2000;4:448-454.
[23]
UNAIDS WHO. AIDS epidemic update: . December 2007 [cited 10 March 2009]; 1-
60]. Available from: http://data.unaids.org/pub/EPISlides/2007/2007_epiupdate_en.pdf
[24]
Mukadi YD, Wiktor SZ, Coulibaly IM, al e. Impact of HIV infection on the
development, clinical presentation and outcome of tuberculosis among children in Cote
d'Ivoire. Aids 1997;11:1151-1158.
[25]
Kiwanuka J, Graham SM, Coulter JBS, et al. Diagnosis of pulmonary tuberculosis in
children in an HIV-endemic area, Malawi. Ann Trop Paeds 2001;21:5-14.
[26]
Jeena PM, Pillay P, Pillay T, Coovadia HM. Impact of HIV-one coinfection on
presentation and hospital-related mortality in children with culture proven pulmonary
tuberculosis in Durban, South Africa. Int J Tuberc Lung Dis 2002;6:672-678.
[27]
Marais BJ, Pai M. New approaches and emerging technologies in the diagnosis of
childhood tuberculosis. Paediatr Resp Rev 2007;8:124-133.
[28]
Marais BJ, Graham SM, Cotton MF, Beyers N. Diagnostic and management
challenges for childhood tuberculosis in the era of HIV. J Infect Dis 2007;196:76-85.
[29]
Walters E, Cotton MF, Rabie H, et al. Clinical presentation and outcome of
Tuberculosis in Human Immunodeficiency Virus infected children on anti-retroviral therapy.
BMC Pediatrics 2008;8:1.
[30]
Zar HJ, Hanslo D, Appoles P. Induced sputum versus gastric lavage for
microbiological confirmation of pulmonary tuberculosis in infants and young children: a
prospective study. Lancet 2005;365:130-134.
[31]
Gie RP, Beyers N, Schaaf HS, Goussard P. The challenge of diagnosing tuberculosis
in children: a perspective from a high incidence area. Paediatr Respir Rev 2004;5:S147-S149.
[32]
Marais BJ, Pai M. Recent advances in the diagnosis of childhood tuberculosis. Arch
Dis Child 2007;92:446-452.
25
[33]
Theart AC, Marais BJ, Gie RP, Hesseling AC, Beyers N. Criteria used for the
diagnosis of childhood tuberculosis at primary health care level in a high-burden, urban
setting. Int J Tuberc Lung Dis 2005;9:1210-1214.
[34]
Marais BJ, Gie RP, Schaaf HS, et al. The spectrum of childhood tuberculosis in a
highly endemic area. Int J Tuberc Lung Dis 2006;10:732-738.
[35]
Hatherill M, Hawkridge T, Zar H, et al. Induced sputum or gastric lavage for
community-based diagnosis of childhood pulmonary tuberculosis? Arch Dis Child
2009;94:195-201.
[36]
Coulter JBS. Diagnosis of pulmonary tuberculosis in young children. Ann Trop
Paediatr 2008;28:3-12.
[37]
Pai M, Flores LL, Pai N. Diagnostic accuracy of nucleic acid amplification tests for
tuberculous meningitis: a systematic review and meta-analysis. Lancet Infect Dis 2003;3:633643.
[38]
Pai M, Riley LW, Colford JMJ. Interferon-gamma assays in the immune diagnosis of
tuberculosis: a systematic review. Lancet Infect Dis 2004;4:761-776.
[39]
Das DK. Fine Needle Aspiration Cytology: its Origin, Development, and Present
Status with Special Reference to a Developing Country, India. Diagn Cytopathol
2003;28:345-351.
[40]
Martin HE, Ellis, E.B. Biopsy by fine needle puncture and aspirations. Ann Surg
1930;92:169-181.
[41]
Zajicek J. Aspiration biopsy cytology. Cytology of supradiaphragmatic organs. Basel:
S Karger 1974.
[42]
Nayak S, Mani R, Kavatkar AN, Puranik SC, Holla VV. Fine-needle aspiration
cytology in lymphadenopathy of HIV-positive patients. Diagn Cytopathol 2003;29:146-148.
26
[43]
Kocjan G, Miller R. The cytology of HIV-induced immunosuppression. Changing
pattern of disease in the era of highly active antiretroviral therapy. Cytopathology
2001;12:281-296.
[44]
Gupta AK, Nayar M, Chandra M. Critical appraisal of fine needle aspiration cytology
in tuberculous lymphadenopathy. Acta Cytol 1992;36:391-394.
[45]
Cruz AT, Starke JR. Clinical manifestations of tuberculosis in children. Paediatric
Respiratory Reviews 2007;8:107-117.
[46]
Rameshkumar K. Tuberculous lymphadenitis in children--role of fine needle
aspiration cytology. J Assoc Physicians India; 1999:976-979.
[47]
Marais BJ, Gie RP, Schaaf HS, et al. The clinical epidemiology of childhood
pulmonary tuberculosis: a critical reading of literature from the pre-chemotherapy era. Int J
Tuberc Lung Dis 2004;8:278-285.
[48]
Marais BJ, Wright CA, Schaaf HS, et al. Tuberculous Lymphadenitis as a Cause of
Persistent Cervical Lymphadenopathy in Children From a Tuberculosis-Endemic Area.
Pediatr Infect Dis J 2006;25:142-146.
[49]
Marias BJ, Gie RP, Schaaf HS, et al. The clinical epidemiology of childhood
pulmonary tuberculosis: a critical reading of literature from the pre-chemotherapy era. Int J
Tuberc Lung Dis 2004;8:278-285.
[50]
Moore SW, Schneider JW, Schaaf HS. Diagnostic aspects of cervical
lymphadenopathy in children in the developing world: a study of 1,877 surgical specimens.
Pediatr Surg Int 2003;19:240-244.
[51]
Marais BJ, Obihara CC, Warren RM, et al. The burden of childhood tuberculosis: a
public health perspective [Review Article]. Int J Tuberc Lung Dis 2005;9:1305-1313.
[52]
Wright CA, Hesseling A, Bamford C, et al. Fine-needle aspiration biopsy: a first-line
diagnostic procedure in paediatric tuberculosis suspects with peripheral lymphadenopathy. Int
J Tuberc Lung Dis 2009;in press.
27
[53]
Graham SM, Coulter JB, Gilks CF. Pulmonary disease in HIV-infected African
children. Int J Tuberc Lung Dis 2001;5:12-23.
[54]
Centres for Disease Controltrol and Prevention. Revised classification system for
human immunodeficiency virus infection in children less than 13 years of age; 1994. Report
No.: MMWR 1994; 43: RR-12.
[55]
Cotton MF, Schaaf HS, Hesseling AC, Madhi SA. HIV and childhood tuberculosis:
the way forward. Int J Tuberc Lung Dis 2004;8:675-682.
[56]
Jeena PM, Coovadia HM, Hadley LG, et al. Lymph node biopsies in HIV infected and
non infected children with persistent lung disease. Int J Tuberc Lung Dis 2000;4:139 -146.
[57]
Zampoli M, Kilborn T, Eley B. Tuberculosis during early antiretroviral-induced
immune reconstitution in HIV-infected children. Int J Tuberc Lung Dis 2007;11:417-423.
[58]
Wakely PE, Kardos TF, Frable WJ. Application of fine needle aspiration biopsy to
pediatrics. Hum Pathol 1988;19:1383-1386.
[59]
Drut R, Drut RM, Pollono D, et al. Fine-Needle Aspiration Biopsy in Paediatric
Oncology Patients. A Review of Experience with 829 Patients. Paediatr Haematol Oncol
2005;27:370-376.
[60]
Thomas JO, Adeyi D, Amanguno H. Fine Needle Aspiration in the Management of
Peripheral Lymphadenopathy in a Developing Country. Diagn Cytopathol 1999;21:159-162.
[61]
Wright CA, Pienaar JP, Marais BJ. Fine needle aspiration biopsy: diagnostic utility in
resource-limited settings. Ann Trop Paediatr 2008; 28:65-70.
[62]
Wright CA, van der Burg M, Geiger D, et al. Diagnosing Mycobacterial lymphadenitis
in children using Fine Needle Aspiration Biopsy: Cytomorphology, ZN staining and
Autofluorescence – making more of less. Diagn Cytopathol 2008;36:245-251.
[63]
Bezabih M, Marium DW, Selassie SG. Fine needle aspiration cytology of suspected
tuberculous lymphadenitis. Cytopathology 2002;13:284-290.
28
[64]
Lau SK, Wei WI, Hsu C, Engzell UCG. Efficacy of fine needle aspiration cytology in
the diagnosis of tuberculous cervical lymphadenopathy. J Laryngol Otol 1990;104:24-27.
[65]
DeMay RM. The Art and Science of Cytopathology. Chicago: ASCP Press 1996.
[66]
Wilson D, Nachega JB, Chaisson RE, Maartens G. Diagnostic yield of peripheral
lymph node needle-core biopsies in HIV-infected adults with suspected smear-negative
tuberculosis. Int J Tuberc Lung Dis 2005;9:220-222.
[67]
Bem C, Patil PS, Elliott AM, et al. The value of wide-needle aspiration in the
diagnosis of tuberculous lymphadenitis in Africa. AIDS 1993;7:1221-1226.
[68]
Bezabih M, Mariam DW, Selassie SG. Fine needle aspiration cytology of suspected
tuberculous lymphadenitis. Cytopathology 2002;13:284-290.
[69]
Wright CA, van Zyl Y, Burgess SM, Blumberg L, Leiman G. Auto fluorescence of
Mycobacteria on Lymph Node Aspirates – A glimmer in the dark? Diagn Cytopathol
2004;30:257-260.
[70]
Marais BJ, Brittle W, Painczyk K, et al. Use of light-emitting diode fluorescence
microscopy to detect acid-fast bacilli in sputum. CID 2008;47:203-207.
[71]
Wright CA, Bamford C, Prince Y, et al. Mycobacterial Transport medium for routine
culture of Fine Needle Aspiration Biopsies. Arch Dis Child 2009;In press.
[72]
Parsons LM, Brosch R, Cole ST, et al. Rapid and simple approach for identification of
Mycobacterium tuberculosis complex isolates by PCR-based genomic deletion analysis. J
Clin Microbiol 2002;40:2339-2345.
[73]
Singh KK, 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.
[74]
van Coppenraet ESB, Lindeboom JA, Prins JM, et al. Real-Time PCR Assay Using
Fine-Needle Aspirates and Tissue Biopsy Specimens for Rapid Diagnosis of Mycobacterial
Lymphadenitis in Children. J Clin Microbiol 2004;42:2644-2650.
29
[75]
Portillo-Gomez L, Murillo-Neri MV, Gaitan-Mesa J, Sosa-Iglesias EG. Nested
polymerase chain reaction in the diagnosis of cervical tuberculous lymphadenitis in Mexican
children. Int J Tuberc Lung Dis 2008;12:1313-1319.
30
Table 1
Summary of the Fine Needle Aspiration Biopsy Technique
Equipment
22 or 23G cutting needles
10ml disposable plastic syringes
Glass cytology slides (ground edges)
Commercial spray fixative or 95% alcohol
Non Sterile gloves
Alcohol swabs
Procedure
Clean skin, immobilize the mass, position needle to access entire mass and avoid
passing through muscles such as sternocleidomastoid
Insert needle firmly and apply no more than 1ml suction, applying constant suction
throughout aspirate
Aspirate moving the needle in a fan like fashion through the mass.
When material is in the hub of the needle, release suction and withdraw needle.
Ask parent or assistant to apply pressure to puncture wound
Preparation of smears
Remove needle from syringe, pull 10 ml of air into syringe and reattach needle.
This air is used to express the material in needle onto a glass slide. Place second glass
slide face down on first, allow material to spread gently, and keeping both slides
together, pull gently apart.
Fix one slide with alcohol and allow the second to air dry
Culture
If liquid culture or transport medium is available, withdraw an aliquot of liquid media
into the syringe and then expel it back into the tube or bottle, thereby using the culture
media to rinse the syringe and needle in a sterile fashion.
31
Table 2
Comparison of different superficial mass lesion sample collection methods to achieve a
definitive tissue diagnosis
Fine Needle
Aspiration
Biopsy
(22G or smaller)
Open surgical
biopsy
Core needle
biopsy
Standard Needle
Aspiration
(greater than
22G)
Suitable for
small lesions
(1x1cm)
Yes
No
Yes
Yes
Entire mass
sampled
Yes, if excision
No
No
Yes
Complications
Risk of
anaesthesia,
hospitalization,
sinus
formation,
infection
Sinus
formation,
infection
Rare
Extremely rare
Cost
Hospitalization, High cost of
theatre time
needle
Minimal
Minimal
Hospitalisation
Yes
No
No
No
Anaesthetic
required
General
Local
Local
None
Time for entire
procedure
1-2 days
30 minutes
30 minutes
10 minutes
Tissue diagnosis
Yes
Yes
Yes
Yes
Microscopy for
organism
Yes
Yes
Yes
Yes
Culture
Yes
Yes
Yes
Yes
Time for initial
result
1-2 days
1-2 days
12-24 hours
(Possible in <1
hr)
12-24 hours
(Possible in <1hr)
32
Table 3
Comparison of methods to confirm a mycobacterial diagnosis in children
Method
Hospitalization
Suitable
Local/general
Equipment /
required
ages
anaesthetic
infrastructure
FNAB
No
All
No
None
Gastric
Yes
All
No
None
Yes
All
No
Nebulizer, suction,
aspirates
Induced
saturation monitor
sputum
Sputum
No
>7-8yrs only No
None
Surgical
Yes
All
Full theatre
biopsy
Yes
33
CHAPTER 3
TB LYMPHADENITIS: A COMMON CLINICAL PROBLEM
Tuberculous Lymphadenitis as a Cause of Persistent Cervical Lymphadenopathy in
Children From a Tuberculosis-Endemic Area
Marais BJ, Wright CA, Schaaf HS, Gie, RP, Hesseling A, Enarson D, Beyers N.
Paediatric Infectious Diseases Journal 2006; 25:142-146.
34
ORIGINAL STUDIES
Tuberculous Lymphadenitis as a Cause of Persistent
Cervical Lymphadenopathy in Children From a
Tuberculosis-Endemic Area
Ben J. Marais, MRCP (Paed), FCP (Paed), MMed,* Colleen A. Wright, MMed, FCPath, FRCPath, FIAC,†
H. Simon Schaaf, MMed, MD (Paed),* Robert P. Gie, FCP (Paed),* Anneke C. Hesseling, MBChB, MSc,*
Don A. Enarson, MD,‡ and Nulda Beyers, FCP (Paed), MSc, PhD*
Background: Cervical lymphadenitis is the most common form of
extrapulmonary tuberculosis in children, although its relative contribution as a cause of persistent cervical adenopathy is not welldocumented. The aim of this study was to determine the relative
contribution of tuberculous lymphadenitis as a cause of persistent
cervical adenopathy in a tuberculosis-endemic setting and to document its clinical presentation at the primary health care level.
Methods: A prospective descriptive study was conducted from
February 2003 through October 200 at 5 primary health care clinics
in Cape Town, South Africa. The study included all children
younger than 13 years presenting with persistent cervical adenopathy to the local primary health care clinic.
Results: A total of 158 children were evaluated of whom 35 (22.2%)
were diagnosed with tuberculous lymphadenitis. Bacteriologic confirmation was achieved in 27 of 35 (77.1%) children; all 35 responded to standard antituberculosis treatment. The majority of
those without tuberculous lymphadenitis (105 of 123, 85.4%) had a
visible superficial lesion in the area drained by the affected nodes. In
children with persistent lymphadenopathy ⱖ2 ⫻ 2 cm, tuberculosis
lymphadenitis was diagnosed in 31 of 33 (93.9%); specificity was
98.4%, sensitivity was 88.6% and the positive predictive value was
93.4%.
Conclusion: Children commonly present with persistent cervical
adenopathy to the primary health care clinic. The use of a simple
clinical algorithm provided an accurate diagnosis of tuberculous
lymphadenitis in the study setting. Fine needle aspirations provided
a rapid and definitive diagnosis in the majority of children and will
have added diagnostic value in settings where alternative diagnoses
are more likely.
Accepted for publication September 13, 2006.
From the *Desmond Tutu TB Center and Department of Paediatrics and
Child Health, Tygerberg Children’s Hospital, and the †Department of
Anatomical Pathology, Tygerberg Hospital, Stellenbosch University,
Cape Town, South Africa; and the ‡International Union Against Tuberculosis and Lung Disease, Paris, France
Dr Marais was supported by Astra Zeneca, the Medical Research Council of
South Africa and the United States Agency for Aid and International
Development. The study is in partial fulfillment of a PhD dissertation.
E-mail bjmarais @sun.ac.za. Reprints not available.
Copyright © 2006 by Lippincott Williams & Wilkins
ISSN: 0891-3668/06/2502-0142
DOI: 10.1097/01.inf.0000199259.04970.d1
142
Key Words: tuberculosis, lymphadenitis, children, persistent
cervical adenopathy
(Pediatr Infect Dis J 2006;25: 142–146)
P
eripheral tuberculous lymphadenitis predominantly involves the cervical lymph nodes1–5 and is the most common form of extrapulmonary tuberculosis in children from
tuberculosis-endemic areas.1,2 However, its relative contribution as a cause of persistent cervical adenopathy in these
communities is not well-documented.
Cervical lymphadenitis, caused by Mycobacterium tuberculosis, is generally considered to have its origin in the
lymphatic spread of organisms from a primary pulmonary
focus, but in a minority of cases it can originate from a
primary focus in the mouth, tonsils, oropharynx or tissues of
the head and neck.6 Other mycobacteria can also cause
cervical lymphadenitis; the relative contribution of different
mycobacteria is influenced by the control of bovine tuberculosis,7 the use of BCG vaccination,8 the presence of environmental mycobacteria and the prevalence of tuberculosis
within a particular setting.7
The diagnosis of tuberculosis in children is often difficult, given that symptoms and signs might be nonspecific,
the collection of bacteriologic specimens problematic and
bacteriologic yields low.9,10 In children with peripheral tuberculous lymphadenitis, however, clinical signs are usually
apparent, and fine needle aspiration (FNA) provides excellent
bacteriologic yields.11–13 Although the diagnostic value of
FNA has been demonstrated in resource-limited settings,11–13
to date it remains underutilized as a routine diagnostic modality in most endemic areas.
The aim of this study was to determine the relative
contribution of tuberculous lymphadenitis as a cause of persistent cervical adenopathy in children from a tuberculosisendemic area and to document its clinical presentation at the
primary health care level.
METHODS
A prospective descriptive study was performed from
February 2003 through October 2004 in Cape Town, the
Western Cape Province, South Africa.
The Pediatric Infectious Disease Journal • Volume 25, Number 2, February 2006
35
The Pediatric Infectious Disease Journal • Volume 25, Number 2, February 2006
Setting. The study was conducted at 5 primary health care
clinics served by one referral hospital. The incidence of all
tuberculosis in Cape Town was 678/100,000,14 and the prevalence of human immunodeficiency virus (HIV) infection among
women attending public antenatal clinics in the Western Cape
Province was 13.1% (95% confidence interval, 8.5–17.7%), in
2003.15 Bovine tuberculosis is well-controlled within the study
communities, and children receive routine neonatal CalmetteGuérin bacillus (BCG) vaccination. The study communities
rarely use private medical services, and children diagnosed with
tuberculosis are routinely referred to the local primary health
care clinic, where supervised antituberculosis treatment is provided free of charge. Pediatric services are extended only to
children younger than 13 years of age.
Study Population. All children (younger than 13 years of age)
who presented with persistent cervical adenopathy; defined as
lymph nodes ⱖ1 ⫻ 1 cm, persisting for ⬎4 weeks despite a
course of oral antibiotics (usually amoxicillin), were referred
to the investigator for evaluation.
Data Collection and Surveillance. The principal investigator
visited each clinic on a weekly basis to screen referred
children, whereas a study nurse recorded children referred
directly to hospital. The areas surrounding the affected cervical lymph nodes were inspected to exclude a visible superficial lesion within their drainage area (visible local cause),
such as impetigo of the scalp, tinea capitis or traction folliculitis. Those with a visible local cause were given appropriate
therapy and instructed to return if the lymph nodes persisted
or increased in size. Surveillance was continued at all 5
clinics and the referral hospital throughout the study period,
and for an additional 3 months after enrollment was stopped,
to document any child who subsequently returned with symptom deterioration or a possible diagnosis of tuberculosis.
Children with no visible local cause of cervical lymphadenopathy received a Mantoux (2 tuberculin units of purified
protein derivative RT23 intradermal) tuberculin skin test
(TST) and a chest radiograph (anteroposterior and lateral).
The Mantoux TST was interpreted as positive if induration
was ⱖ10 mm. Two independent experts evaluated the chest
radiographs for intrathoracic signs of tuberculosis. Symptoms
and signs as well as possible tuberculosis contact were
documented. Children diagnosed with tuberculous lymphadenitis were offered a rapid HIV test (Determine HIV 1/2;
Abbott) after appropriate counseling. Standard 3-drug antituberculosis treatment was initiated,16 and response to antituberculosis therapy was monitored after a period of 3 months.
In the group not diagnosed with tuberculous lymphadenitis,
symptom resolution in the absence of antituberculosis therapy
was monitored.
Case Definition. Tuberculous lymphadenitis was defined as
cervical lymphadenopathy ⱖ1 ⫻ 1 cm persisting for ⬎4
weeks despite a course of oral antibiotics, in the presence of
at least 1 of the following criteria: (1) bacteriologic confirmation: isolation of M. tuberculosis from a lymph node, or
microscopically visible acid-fast or autofluorescent bacilli
associated with amorphous necrosis on cytology; or (2) clinical diagnosis: significant therapeutic response (lymph node
size decreased from ⱖ2 ⫻ 2 cm to ⬍1 ⫻ 1 cm after 3 months
of standard antituberculosis treatment). All cultures positive
© 2006 Lippincott Williams & Wilkins
Tuberculous Lymphadenitis
for M. tuberculosis complex were routinely speciated by
polymerase chain reaction to differentiate M. tuberculosis
from Mycobacterium bovis and M. bovis BCG.17
FNA and/or Biopsy. Of the children enrolled, only those with
a positive TST (ⱖ10 mm) and/or a suggestive chest radiograph and/or a cervical mass ⱖ2 ⫻ 2 cm were referred to the
hospital for FNA. FNA was performed in the pediatric outpatient department with sedation and pain relief (chloral
hydrate 20 –50 mg/kg and paracetamol 10 –20 mg/kg, administered orally 30 – 60 minutes before the procedure). The
largest discrete node was identified by careful palpation of the
cervical mass. Two needle passes were performed with a
23-gauge needle attached to a 10-mL syringe, applying 2 mL
of suction. Aspirated material was placed on standard microscope slides, thinly smeared and air-dried or fixed with
commercial spray fixative for Giemsa and Papanicolou stains,
before flushing the syringe and needle directly into a MGIT
(Becton Dickinson, Cockeysville, MD) mycobacterial culture
bottle. Microscopic evaluation included cytology, acid-fast
staining and autofluorescence.18 Excision biopsies were occasionally performed if children presented with persistent
cervical adenopathy to the referral hospital without a letter
from the investigator or if the FNA was inconclusive.
Statistical Analysis. Descriptive analysis was conducted with
SPSS (version 11.5). The sensitivity and specificity of a
simple clinical algorithm were calculated with the use of
either bacteriologic confirmation and/or the clinical case
definition (see case definition) as the “reference” test.
Parents gave written informed consent for study participation and ethics approval was obtained from the Institutional Review Board of Stellenbosch University, the City of
Cape Town Health Department and local health committees.
RESULTS
Persistent Cervical Adenopathy. Of the 167 children who
were identified with persistent cervical adenopathy, 9 (5.4%)
did not return to the clinic for evaluation by the investigator
(Fig. 1). Of the 158 children evaluated, 53 had no visible
cause of whom 40, with either a positive TST or a cervical
mass ⱖ2 ⫻ 2 cm, were referred to hospital to establish a
histologic diagnosis. None of the 13 TST-negative children
had radiographic or other clinical signs indicative of possible
tuberculosis, and all children judged not to have tuberculosis
on clinical grounds showed symptom resolution in the absence of antituberculosis treatment.
Table 1 summarizes the demographics and etiology of
persistent cervical adenopathy in the 158 children who were
evaluated. The majority of children (105, 66.5%) had lymph
nodes ⬍2 ⫻ 2 cm with a visible local cause; 28 returned for
reevaluation within 1 month; the lymphadenopathy resolved
in 18 and decreased to ⬍1 ⫻ 1 cm in 9. Multiple discrete
lymph nodes, 1–2 cm in diameter, remained in 1 child who
was TST-negative and asymptomatic. The lymphadenopathy
showed slow resolution, decreasing to ⬍1 ⫻ 1 cm within 3
months. None of the children with a visible local cause
returned to the clinic with symptom deterioration or with a
possible tuberculosis diagnosis from any hospital, during the
study period.
143
36
Marais et al
The Pediatric Infectious Disease Journal • Volume 25, Number 2, February 2006
FIGURE 1. Flow diagram of children referred with persistent cervical adenopathy. Not TB indicates symptom resolution in the
absence of antituberculosis chemotherapy. TB, bacteriologic confirmation: isolation of M. tuberculosis from a lymph node, or microscopically visible acid-fast or autofluorescent bacilli associated
with caseating necrosis on cytology or clinical diagnosis; significant
therapeutic response (lymph node size decreased from ⱖ2 ⫻ 2
cm to ⬍1 ⫻ 1 cm after 3 months of standard antituberculosis
treatment); Not evaluated, did not return to the clinic for evaluation by the investigator; Rx response, clinical diagnosis together
with significant therapeutic response in the absence of FNA or
biopsy-based diagnosis; ⴱclassified as not TB: 1 chronic inflammatory process diagnosed after excision, 1 nonacute bacterial abscess
diagnosed after incision and drainage.
TABLE 1. Demographics and Etiology of Persistent
Cervical Adenopathy in Children (n ⫽ 158)
No. of Instances
Demographics
Gender
Male
Female
Age groups
⬍5 yr
5–9 yr
ⱖ10 yr
Etiology
Visible local cause
Bacterial infection (crusted impetigo)
Tinea capitis (with secondary infection)
Traction folliculitis
Otitis externa
No visible local cause
Tuberculous lymphadenitis
Reactive nodes†
Nonspecific inflammation
Nonacute bacterial abscess
Malignancy
69 (43.7)*
89 (56.3)
93 (58.9)
51 (32.2)
14 (8.9)
105 (66.5)
26 (16.5)
34 (21.5)
44 (27.8)
1 (0.6)
53 (33.5)
35 (22.2)
13 (8.2)
4 (2.5)
1 (0.6)
0
*Numbers in parentheses, percent.
Cervical mass ⬍2 ⫻ 2 cm, tuberculin skin test-negative and natural symptom
resolution.
†
144
Of the 53 (33.5%) children without a visible local cause,
tuberculous lymphadenitis was diagnosed in 35 (66.0%). Of
those with a cervical mass ⱖ2 ⫻ 2 cm, tuberculous lymphadenitis was diagnosed in 31 of 33 (93.9%). Of the 2 remaining
children, 1 had an unidentified chronic inflammatory process
diagnosed through excision biopsy, and the other developed a
chronic bacterial abscess after repeat courses of oral antibiotics,
requiring eventual incision and drainage. Two children presented with acute bacterial adenitis; both had a history of
persistent lymph node enlargement that preceded the acute
event. Both these children were TST-positive and were
diagnosed with secondary bacterial infection, as M. tuberculosis was cultured after incision and drainage. One had
signs suggestive of tuberculosis on the chest radiograph
and reported a prolonged cough and night sweats, whereas
the other had no additional suspect symptoms or signs
apart from the cervical mass and a positive TST.
Tuberculous Lymphadenitis. Of the 35 children diagnosed with
tuberculous lymphadenitis (Fig. 1), bacteriologic confirmation
was achieved in 27 (77.1%) children. Of the 8 children without
bacteriologic confirmation, 7 failed to attend the referral hospital, and 1 refused permission for FNA. All 8 had a TST
response ⱖ15 mm and showed excellent response to standard
antituberculosis treatment. In the 27 children with bacteriologic confirmation; FNA was performed in 21 children and
formal biopsies in 6. No cases caused by M. bovis, M. bovis
BCG, infection with environmental mycobateria such as
Mycobacterium avium-intracellulare complex or Mycobacterium scrofulaceum or malignancies were identified.
Of the 21 children with tuberculous lymphadenitis in
whom FNA was performed, 16 of 21 (76.2%) were acid-fast
or autofluorescent smear-positive, which allowed rapid and
definitive diagnosis. Culture confirmation was achieved in 19
of 21 (90.5%), of whom 2 had no microscopic features
indicative of tuberculosis. One of the 2 children who were
culture-negative had received antituberculosis treatment before FNA was performed. Either a positive culture or typical
microscopic features were present in all 21 cases. One of the
2 children, who had no microscopic features indicative of
tuberculosis on FNA, had an excision biopsy performed that
established the diagnosis, before the FNA culture result
became known. No immediate complications relating to the
FNA procedure, apart from minimal bleeding, were noted.
No long term complications, such as sinus formation, were
recorded during the 3-month follow-up period.
Table 2 reflects the lymph node characteristics and
associated findings in the 35 children diagnosed with tuberculous lymphadenitis. Using the clinical algorithm of a persistent cervical mass ⱖ2 ⫻ 2 cm, without a visible local
cause or response to antibiotics, accurately identified children
with tuberculous lymphadenitis; sensitivity was 88.6%, specificity was 98.4% and the positive predictive value of this
clinical algorithm was 93.4%.
Tuberculous lymphadenitis occurred in children of all
ages, except in infants. In 18 (51.4%), lymph nodes occurred
in the anterior triangle, with involvement of multiple regions
in 5 (14.3%). Other regions involved were: posterior triangle,
8 (22.9%); submandibular, 2 (5.7%); and supraclavicular, 2
(5.7%). Lymph nodes occurred more regularly on the right
© 2006 Lippincott Williams & Wilkins
37
The Pediatric Infectious Disease Journal • Volume 25, Number 2, February 2006
TABLE 2. Clinical Characteristics of Children With
Tuberculous Lymphadenitis (n ⫽ 35)
No. of Instances
Lymph node characteristics
Persistence (present for ⬎4 wk, no response to
antibiotics)
Size†
⬍2 ⫻ 2 cm
(2– 4) ⫻ (2– 4) cm
⬎4 ⫻ 4 cm
Character
Single
Multiple
Discreet
Matted
Solid
Fluctuant
Without secondary bacterial infection
With secondary bacterial infection (red
and warm)
Associated findings
Tuberculin skin test
0 mm
1–9 mm
ⱖ10 mm
ⱖ15 mm
Mean response 19.1 mm (standard deviation
2.9 mm)
Constitutional symptoms
Any symptom
Fever
Cough
Night sweats
Fatigue‡
Failure to thrive§
Chest radiograph
Suggestive of tuberculosis
Lymph node disease
Uncomplicated
With airway compression
With parenchymal consolidation
35 (100)*
4 (11.4)
25 (71.5)
6 (17.1)
The response to standard antituberculosis therapy was
good, and most children showed considerable improvement,
with reduction in lymph node size to ⬍1 ⫻ 1 cm after 3
months of standard antituberculosis treatment. Three (8.6%)
children deteriorated initially; 1 developed upper airway
obstruction that required surgical lymph node enucleation and
the addition of corticosteroids. All 3 children showed good
response after completing 6 months of standard antituberculosis therapy, although a cervical mass of ⬎1 ⫻ 1 cm remained in 1 child.
5 (14.3)
14 (40.0)
16 (45.7)
28 (80.0)
5 (14.3)
2 (5.7)
2 (5.7)
0
33 (94.3)
32 (91.4)
21 (60.0)
7 (20.0)
9 (25.7)
8 (22.8)
19 (54.3)
10 (28.6)
13 (37.1)
8 (22.8)
1 (2.9)
4 (11.4)
*Numbers in parentheses, percent.
Transverse diameter of the largest cervical mass.
Less playful and active since the mass was first noted.
§
Crossing at least 1 centile line in the preceding 3 months or having lost ⬎10% of body
weight (minimum, 1 kg) over any time interval.
†
‡
side: right, 19 (54.3%); left, 13 (37.1%); bilateral, 3 (8.6%).
Most lymph nodes were multiple, either discreet or matted
(30, 85.7%), either discreet or matted, and solid (28, 80.0%).
All lymph nodes were nontender, either discreet or matted,
except 2 with secondary bacterial infection. A sinus developed in 2 children, one before presentation and the other after
surgical incision and drainage.
Associated findings in children with tuberculous lymphadenitis included a positive TST in 33 of 35 (94.3%). Two
(5.7%) children had a nonreactive TST; 1 was HIV-infected,
the other was severely ill with coexistent pulmonary tuberculosis. The TST in the latter patient measured 16 mm when
repeated after 3 months of antituberculosis treatment. Twenty-nine (82.9%) children were tested for HIV, and 1 (3.5%)
was HIV-infected. Contact with an adult index case was
reported in 19 (54.3%), whereas in a further 2 cases, adults
with suspicious symptoms were subsequently diagnosed with
tuberculosis. Constitutional symptoms were present in 21
(60.0%) children, and the chest radiograph was suggestive of
primary pulmonary tuberculosis in 13 (37.1%).
© 2006 Lippincott Williams & Wilkins
Tuberculous Lymphadenitis
DISCUSSION
Children commonly presented with persistent cervical
adenopathy to primary health care clinics in the study setting.
The majority of these children had visible local lesions such
as crusted impetigo, infected tinea capitis or traction folliculitis. The hair of most small girls in this community is tightly
braided, which often leads to irritation and bacterial infection
of the hair follicles. This also explains the overrepresentation of girls in the study population. It is understandable that
none of these conditions responded to a short course of oral
antibiotics, in the absence of etiology-specific systemic
and/or local treatment.
A simple clinical algorithm that identified children with
persistent (⬎4 weeks) cervical lymphadenopathy, no visible
local cause or response to antibiotics and a cervical mass
ⱖ2 ⫻ 2 cm showed excellent diagnostic accuracy within the
study setting. The addition of a positive TST may be of value
in settings where infection with M. tuberculosis is less common or where conditions other than tuberculosis, such as
malignant lymph node involvement or other chronic infections, are more common and may present with a similar
clinical picture. The fact that accurate clinical diagnosis is
possible at the primary health care level might allow the
initiation of antituberculosis treatment without hospital referral, which should improve access to care for children in
extremely resource-limited settings. However, regular follow-up (at least monthly) is essential so that children who
do not respond to standard antituberculosis treatment are
referred as soon as possible to establish a definitive diagnosis. The value of this clinical approach requires further
evaluation.
FNA proved to be a robust and simple technique, which
provided an excellent bacteriologic yield. No significant side
effects were noted. FNA provided a definitive bacteriologic
diagnosis with the ability to speciate M. tuberculosis complex
and to perform drug susceptibility testing. A definitive diagnosis is always desirable, but the diagnostic value of FNA
will be even greater in settings where a clinical diagnosis is
expected to be less accurate, such as areas where malignant
lymph node involvement and/or other chronic infections are
more common.11,12
The lymph node characteristics and constitutional symptoms recorded in this study correlate well with findings from
Papua New Guinea and India.3,4,19 The most distinctive features were the chronic persistent course, the lymph node size
(ⱖ2 ⫻ 2 cm) and the involvement of multiple, discrete or
matted, lymph nodes. The study also emphasizes the value of
145
38
The Pediatric Infectious Disease Journal • Volume 25, Number 2, February 2006
Marais et al
clinical follow-up and serial measurement of enlarged lymph
nodes if the diagnosis is not apparent.20
Only 1 child with tuberculous lymphadenitis was HIVinfected. This reflects the low HIV prevalence among children in the study setting but might also illustrate a decreased
tendency in HIV-infected children to develop peripheral
tuberculous lymphadenitis, as suggested in a comparative
study from Zambia.21 FNA may have increased diagnostic
value in HIV-infected children, because the relative contribution of tuberculous lymphadenitis to persistent cervical
adenopathy in this group is expected to be smaller.
It is interesting that no environmental mycobacteria
were isolated, although it is reported to be the most common
cause of persistent cervical adenopathy in the developed
world.22 Paucity of disease caused by environmental mycobacteria has also been described in other tuberculosis-endemic countries, such as India.23,24 This may result from the
protective effect afforded by routine neonatal BCG vaccination,8 whereas natural infection with M. tuberculosis might
provide additional protection against disease caused by environmental mycobacteria.
An important study limitation is that bacteriologic confirmation was achieved in only 77.1% of the study patients, as it
was not attempted in 8 children. However, we are confident that
the diagnosis was accurate, given the fact that these 8 children all
had a TST response ⱖ15 mm and showed excellent clinical
response to standard antituberculosis treatment.
In conclusion, the use of a simple clinical algorithm
identified tuberculous lymphadenitis with a high degree of
accuracy in the study setting, whereas FNA provided a rapid
and definitive diagnosis in the majority of children.
ACKNOWLEDGMENTS
We thank the primary health care clinics involved, Dr
Ivan Toms (City of Cape Town Health Department), the
patients and their parents for their kind assistance.
REFERENCES
1. Coovadia HM. Tuberculosis in children. In: Coovadia HM, Benatar SR,
eds. A Century of Tuberculosis. Cape Town, South Africa: Oxford
University Press; 1991:91–100.
2. Seth V. Antituberculous therapy in children. Indian J Pediatr. 1986;53:
174 –198.
3. McMaster P, Ezeilo N, Freisen H, Pomat N, Vince JD. Ten-year
experience with paediatric lymph node tuberculosis in Port Moresby.
J Trop Pediatr. 2001;47:160 –164.
4. Seth V, Kabra SK, Jain Y, et al. Tubercular lymphadenitis: clinical
manifestations. Indian J Pediatr. 1995;62:565–570.
146
5. Reddy MP, Moorchung N, Chaudrey A. Clinico-pathological profile of
paediatric lymphadenopathy. Indian J Pediatr. 2002;69:1047–1051.
6. Miller FLW, Cashman JM. The natural history of peripheral tuberculous
lymphadenitis associated with a visible primary focus. Lancet. 1955;1:
1286 –1289.
7. Grzybowski S, Allen EA. History and importance of scrofula. Lancet.
1995;346:1472–1474.
8. Romanus V. Tuberculosis in BCG immunized and unimmunized
children in Sweden: a ten-year evaluation following the cessation of
BCG immunization of the newborn in 1975. Pediatr Infect Dis J.
1987;6:272–280.
9. Eamranond P, Jamarillo E. Tuberculosis in children: reassessing the
need for improved diagnosis in global control strategies. Int J Tuberc
Lung Dis. 2001;5:594 – 603.
10. Starke JR. Pediatric tuberculosis: time for a new approach. Tuberculosis.
2003;83:208 –212.
11. Handa U, Palta A, Mohan H, Punia RP. Fine needle aspiration diagnosis
of tuberculous lymphadenitis. Trop Doct. 2002;32:147–149.
12. Lau S, Ignace W, Kwan S, Yew W. Combined use of fine needle
aspiration cytologic examination and tuberculin skin test in the diagnosis
of cervical tuberculous lymphadenitis. Arch Otolaryngol Head Neck
Surg. 1991;117:87–90.
13. Thomas J, Adeyi AO, Olu-eddo AO, Nwachokor N. Fine needle aspiration cytology in the management of childhood palpable masses: Ibadan
experience. J Trop Pediatr. 1999;45:378.
14. Cape Town TB Control. Progress Report 1997–2003. Available at
www.hst.org.za/publications/618.
15. Department of Health. National HIV and Syphilis antenatal sero-prevalence survey 2003. Available at www.doh.gov.za/docs/reports/2003/
hiv/p1-23.
16. Department of Health. The South African Tuberculosis Control Programme: Practical Guidelines. 2000:32–37.
17. de Wit D, Steyn L, Shoemaker S, Sogin M. Direct determination of
Mycobacterium tuberculosis in clinical specimens by DNA amplification. Clin Microbiol. 1990;28:2437–2441.
18. Wright CA, van Zyl Y, Burgess SM, Blumberg L, Leiman G.
Mycobacterial autofluorescence in Papanicolaou-stained lymph node
aspirates: a glimmer in the dark? Diagn Cytopathol. 2004;30:257–
260.
19. Narang P, Narang R, Narang R, et al. Prevalence of tuberculous
lymphadenitis in children in Wardha district, Maharashtra State, India.
Int J Tuberc Lung Dis. 2005;9:188 –194.
20. Knight PJ, Mulne AF, Vassy LE. When is lymph node biopsy indicated
in children with enlarged peripheral nodes? Pediatrics. 1982;69:391–
396.
21. Chintu C, Bhat G, Luo C, et al. Seroprevalence of human immunodeficiency virus type 1 infection in Zambian children with tuberculosis.
Pediatr Infect Dis J. 1993;12:499 –504.
22. Lai KK, Stottmeier KD, Sherman IH, McCabe WR. Mycobacterial
cervical lymphadenopathy. JAMA. 1984;251:1286 –1288.
23. Krishnaswami H, Koshi G, Kulkarni KG, Job CK. Tuberculous lymphadenitis in South India: a histopathological and bacteriological study.
Tubercle. 1972;53:215–220.
24. Aggarwal P, Wali JP, Singh S, et al. A clinico-bacteriological study
of tuberculous lymphadenitis. J Assoc Physicians India. 2001;49:
808 – 812.
© 2006 Lippincott Williams & Wilkins
39
CHAPTER 4
DIAGNOSTIC UTILITY AND FEASIBILITY OF FINE
NEEDLE ASPIRATION BIOPSY
Fine needle aspiration biopsy: diagnostic utility in resource-limited settings
C A Wright, J P Pienaar,Ben J Marais
Annals of Tropical Paediatrics. 2008; 28:65-70.
Fine Needle Aspiration Biopsy - A First Line Diagnostic Procedure in
PediatricTuberculosis suspects with Peripheral Lymphadenopathy?
Colleen A Wright, MD FRCPath; Anneke C Hesseling, MD, PhD ; Colleen Bamford, MD,
FCPath; Steven Michael Burgess, PhD; Rob Warren, PhD; and Ben J Marais, MD, PhD
International Journal of Tuberculosis and Lung Disease-in press
40
Annals of Tropical Paediatrics (2008) 28, 65–70
Fine needle aspiration biopsy: diagnostic utility in
resource-limited settings
C. A. WRIGHT, J. P. PIENAAR & B. J. MARAIS*
Discipline of Anatomical Pathology, Department of Pathology and NHLS Tygerberg and *Department of
Paediatrics & Child Health and Ukwanda Centre for Rural Health, University of Stellenbosch, Tygerberg,
South Africa
(Accepted November 2007)
Abstract
Background: Little information is available on the value of fine needle aspiration biopsy (FNAB) in routine
paediatric practice in resource-limited settings.
Aim: To provide an overview of all paediatric FNAB samples received at Tygerberg Hospital, Cape Town, South
Africa over a 3-year period, including the determinants of sample adequacy and the diagnoses.
Methods: Samples were analysed from three locations: Tygerberg Hospital (TBH) where pathologists performed all
the procedures, surrounding clinics where aspirates were mostly performed by doctors with no formal training in
FNAB technique, and Queen Elizabeth Hospital, Blantyre, Malawi where FNABs were performed by trained nurse
aspirators.
Results:A total of 830 aspirates were reviewed: 464 (56%) from TBH, 264 (32%) from local clinics and 102 (12%)
from Blantyre. The main diagnoses at TBH were mycobacterial infection (31%), normal/reactive tissue (27%) and
malignancy (14%); malignancy dominated (74%) in the select group from Blantyre. Sample adequacy rates were
similar between pathologists and nurse aspirators [399/464 vs 82/102, odds ratio (OR) 1.4, 95% confidence
interval (CI) 0.8–2.6]. Results were significantly better in the group who received formal training (TBH and
Malawi) than in the clinics where clinicians had no formal training (481/566 vs 171/264, OR 3.1, 95% CI 2.2–
4.4).
Conclusions: FNAB provides a definitive tissue diagnosis in the majority of patients. Well-trained nurse aspirators
perform as well as pathologists, indicating the feasibility of FNAB in resource-limited settings.
Introduction
Fine needle aspiration biopsy (FNAB) has
been used as a diagnostic modality for more
than a century. It was initially applied to
infective and inflammatory lesions only, but
the advent of cytological investigation
(around 1925) expanded its diagnostic
application to confirm or exclude malignancy also.1 FNAB is well established as
the diagnostic procedure of choice for
Reprint requests to: Professor C. A. Wright, Discipline
of Anatomical Pathology, PO Box 19063, Tygerberg
7505, South Africa. Email: cawr@sun.ac.za; fax: z27
21 938 6559.
# 2008 The Liverpool School of Tropical Medicine
DOI: 10.1179/146532808X270707
superficial mass lesions.2–4 However, few
data are available on its routine use in
paediatric practice, especially in resourcelimited settings.
FNAB provides a simple, fast, costeffective and safe method of obtaining a
definitive tissue diagnosis.3–7 In experienced
hands, a high level of diagnostic accuracy is
possible, averting the need for a formal
biopsy and all its associated risks.8 Apart
from providing a cytological diagnosis,
FNAB also provides tissue for ancillary
tests, such as microbiological cultures, flow
cytometry, cytogenetics, electron microscopy and polymerase chain reaction
(PCR).2,5
41
66
C. A. Wright et al.
If the procedure is performed according
to established protocols, the disadvantages
are minimal.9 Potential complications
include pain, bleeding, infection, vasovagal
reactions, haemothorax, pneumothorax, air
embolism, seeding along the needle tract
and perforation of organs.9 As a general
rule, the rate of complications rises with use
of larger calibre needles and deep organ
aspiration. Contra-indications for FNAB are
few but include an abnormal clotting profile
(especially for deep organ aspirations) and
suspected vascular lesions, and aspiration in
the region of the thorax is best avoided in
patients with cardiovascular and/or respiratory dysfunction.9 The risk associated
with the aspiration of superficial masses
using small gauge needles (22G or less) is
minimal.2,9
Previously at Tygerberg Hospital, clinicians without formal training in the correct
technique performed routine FNAB, but
since July 2001 trained pathologists perform
all paediatric fine needle aspirates. The
study aims were to provide an overview of
all diagnoses made during the 1st 3 years of
this new policy and to investigate possible
determinants of sample adequacy.
Methods
Study population
This
retrospective
descriptive
study
reviewed all paediatric FNAB samples
received by the cytology laboratory at
Tygerberg Hospital (TBH) from 1 July
2001 to 30 June 2004. These included
samples received from TBH and surrounding clinics and specimens couriered from
Queen Elizabeth Hospital, Blantyre,
Malawi. The cases referred from Malawi
were all enrolled in a paediatric oncology
study and represent a very select group.
Children were recruited into this study and
underwent FNAB if they presented with a
mass lesion suspected of being Burkitt’s
lymphoma.10
Fine needle aspiration biopsy
FNAB were usually performed in the clinic,
in the paediatric outpatient department or in
the ward. Children received sedation and
pain relief (usually chloral hydrate 20–50 mg/
kg and paracetamol 10–20 mg/kg) orally 30–
60 minutes before the procedure. Before
performing the aspiration, the largest discrete
mass was identified by careful palpation,
(usually using a 23G-needle attached to a
10-ml syringe and applying 2 ml of suction).
Aspirated material was placed on standard
microscope slides, thinly smeared and airdried or fixed with commercial spray fixative
for Giemsa and Papanicolaou stains. If
tuberculosis was clinically suspected, the
needle and syringe were rinsed in mycobacterial growth tube indicator (MGIT, Beckton
Dickinson, USA) medium for mycobacterial
culture. All mycobacterial isolates were
identified as Mycobacterium tuberculosis,
Mycobacterium bovis BCG or non-tuberculous
mycobacteria (NTM) by polymerase chain
reaction (PCR) testing.
Data collection and analysis
Cytology results were captured in a computerised database. We reviewed the results of
all FNAB samples received during the 3year study period from patients (16 years
of age. Relevant information including
patient demographics, specimen characteristics and final diagnosis was transferred to a
Microsoft Excel spreadsheet. Descriptive
statistical analysis was performed using
Statistica 7.
Written, informed consent to the FNAB
procedure was obtained from the parent
or legal guardian. Ethics clearance was
obtained from the institutional review board
of the University of Stellenbosch.
Results
A total of 830 aspirates were reviewed, 464
(56%) from TBH, 264 (32%) from local
42
Fine needle aspiration biopsy
clinics and 102 (12%) from Malawi. In
general, the age and gender distribution was
uniform (406, 51% female), but Malawi had
more males (63% male vs 37% female) and
the clinics had more females (57% female vs
43% male).
Peripheral lymph nodes were most frequently aspirated (376, 45%); of the lymph
nodes specified, 70% were cervical and 30%
axillary. Aspiration sites for the remaining
454 samples were distributed as follows:
head and neck (124, 15%), chest and
abdomen (98, 12%), breast (66, 8%), site
not specified (166, 20%). All 66 breast
tissue samples were collected from pubertal
girls and came from the clinics, constituting
25% of clinic samples. In Malawi where all
the children were clinically suspected of
having Burkitt’s lymphoma, the bulk of
aspirates (44, 43%) came from the head
and neck area, 28 (27%) from the chest,
abdomen or pelvis and in 30% of cases the
site was not specified.
The cytological diagnosis was assessed in
conjunction with the clinical and radiological findings and all clinically or cytologically
suspected lymphomas at TBH were aspirated for flow cytometry. Where appropriate, immunocytochemistry was performed
on neoplasms. All solid neoplasms were
managed according to SIOP (International
Society for Paediatric Oncology) protocols,
and histology was done after chemotherapy.
Where there was any uncertainty as to the
precise diagnosis after FNAB, the mass was
excised or biopsied prior to chemotherapy.
All cases of suspected mycobacterial
infection received bedside inoculation for
mycobacterial culture and subsequent
speciation.
All aspirates from Malawi were referred
after clinical assessment as Burkitt’s lymphoma and, when confirmed by cytology,
patients were commenced on limited chemotherapy according to the study protocols
and response to treatment was carefully
monitored.
The most common diagnostic category
was normal/reactive tissue, followed by
67
mycobacterial infection and lymphoma, in
particular Burkitt’s lymphoma. However,
the distribution of diagnostic categories
showed marked variation between the different centres. Mycobacterial disease was
the most common diagnosis in children
from TBH and Burkitt’s lymphoma in the
select group from Malawi. Among the
children with mycobacterial infection, M.
bovis BCG was identified in 20 (22.5%)
cases. All these children had right-sided
axillary lymph node enlargement and were
,2 years of age (average age ,1 y). M.
tuberculosis was identified in 67 (75.3%)
cases, mostly from cervical lymph nodes,
and was equally prevalent in all age groups
(average age 5.7 y). In Malawi, Burkitt’s
lymphoma was seen predominantly between
the ages of 2 and 12 years (average age 6.3),
corresponding to the age of maximum
prevalence of endemic Burkitt’s lymphoma.11 No significant seasonal or other
time-related variation was recorded on
reviewing the temporal profile of cases with
TB or Burkitt’s lymphoma.
Of all aspirates received, 464 (56%) were
performed by pathologists (TBH), 264
(32%) by clinicians (clinics) and 102
(12%) by nurse aspirators (Malawi).
Pathologists and nurse aspirators received
formal training in FNAB technique but few
of the clinicians. Fig. 1 indicates the adequacy of samples received; in total, 174
(21%) samples were regarded as inadequate. The proportion of inadequate samples varied widely between centres, TBH
14%, Malawi 20% and clinics 35%. This
reflects differences between the aspirator
groups, adequate samples being 86% in the
pathologist group, 80% in the nurse aspirator group and 65% in the untrained clinician
group. Sample adequacy rates were similar
between pathologists and nurse aspirators
(399/464 vs 82/102; OR 1.4, 95% CI 0.8–
2.6) but were significantly better in the
group who received formal training (TBH
and Malawi) than in the clinics where
clinicians did not receive formal training
(481/566 vs 171/264, OR 3.1, 95% CI
43
68
C. A. Wright et al.
FIG. 1. Adequacy rates of FNAB by aspirator: pathologists 86%, untrained clinicians 65%, nurse aspirators 80%
(& inadequate, % adequate).
2.2–4.4). The sample adequacy rate also
differed according to the number of passes
performed, single pass 75% and more than
one pass 85% (85% vs 75%, OR 1.9, 95%
CI 1.3–2.8).
Discussion
FNAB is widely used as a first-line diagnostic procedure for the diagnosis of mass
lesions in adults, but it is used less frequently in children, especially so in
resource-limited settings where it has the
greatest diagnostic potential. This study
demonstrates that FNAB is a feasible option
in resource-limited settings; well trained
nurse aspirators performed as well as trained
pathologists and the procedure can be done
on an outpatient basis.
Nurse aspirators from Malawi who
attended a short training course at TBH
and received clear operating procedure
guidelines achieved an adequacy rate of
80%, despite the fact that multiple needle
passes were restricted. The technique shown
was subsequently implemented within the
units. This was a highly select patient group
with clinically diagnosed Burkitt’s lymphoma
who were enrolled in a trial that investigated
cost-effective management of Burkitt’s lymphoma in resource-limited countries.10 In
the absence of a local cytopathologist, the
slides were fixed and sent to TBH for review.
This trial demonstrated that diagnostic slides
can be transported easily and safely using
standard courier services, enabling cytopathologists to provide a long-distance
diagnostic service for countries without
cytopathology services.
The diagnostic value of FNAB is completely dependent on the quality of the
sample presented to the cytopathologist.
Optimal FNAB technique together with
adequate slide preparation and preservation
is taught poorly, if at all, to medical students
in South Africa. The technique is usually
acquired by observing and assisting senior
clinicians, most of whom receive no formal
training, resulting in the propagation of
incorrect techniques and a high percentage
of inadequate samples. Samples are frequently bloody, inadequately spread and
poorly fixed, making it impossible for the
pathologist to interpret the smears. The
resulting non-diagnostic reports promote
the misconception that FNAB is a technique
with a poor diagnostic yield.
Before January 2001, FNAB was not
widely used by clinicians at TBH. This
was partly due to inadequate training in
FNAB technique and under-recognition of
the valuable contribution that FNAB could
make to the management of infective and
neoplastic lesions in children. Since 2001,
cytopathologists have provided the FNAB
service at TBH, and the diagnostic use and
yield of FNAB in children has increased; an
outreach programme has been established
offering short practical tutorial sessions to
44
Fine needle aspiration biopsy
interested clinicians, combined with
follow-up of subsequent aspirates received,
providing positive feedback and constructive
advice.
The advanced training and experience of
the cytopathologists might partly explain the
excellent results achieved, although nurse
aspirators in Malawi who received training in
the correct technique achieved comparable
results. The performance of more than one
needle pass also improved sample adequacy
and diagnostic yield. In total, more than one
needle pass was performed in 40% of
patients, 59% at TBH, 18% in the clinics
and 11% in Malawi. In Malawi, multiple
needle passes were restricted in an attempt to
minimise the risk of needle tract spread, a
precaution recommended in children suspected of having a malignant neoplasm.
This lack of training is not unique to
developing countries. In a study of breast
aspirates, Ljung showed that physicians
trained in FNAB technique missed significantly fewer malignant lesions than physicians who had received no training (2% vs
25%, p,0.0001).12 This was not affected by
the number of aspirates performed by
individual aspirators.
Few studies have reported on the value
and limitations of FNAB in routine paediatric practice. In the USA, Wakely reported a
sensitivity and specificity rate of 97% and a
positive predictive value of 95% throughout
the entire age range of infancy and childhood.3 Handa showed FNAB to be an
excellent procedure for the triage of patients
with significant lesions requiring treatment
or referral to specialised services in India.8 A
paediatric oncology study in the USA
demonstrated that FNAB is an excellent
tool for the diagnosis of primary and
recurrent malignant neoplasms.6 These
results were supported by a South African
study of paediatric oncology patients where
FNAB showed a sensitivity of 96.1% and a
specificity of 100%.13
FNAB is used increasingly for the diagnosis of infectious disease, particularly
tuberculosis.14,15 Poor countries bear the
69
brunt of paediatric tuberculosis,16 necessitating optimal use of limited resources. It
has been reported that cervical lymphadenitis is the most common extra-thoracic
manifestation of paediatric tuberculosis in
India and South Africa,17–19 but differentiation from other infections such as M. bovis
BCG, non-tuberculous mycobacteria, fungal infections or malignancies such as
Burkitt’s lymphoma or Kaposi’s sarcoma is
important.20,21 FNAB is a minimally invasive procedure that offers a definitive tissue
diagnosis and can be performed by well
trained nurses on an outpatient basis. It has
particular diagnostic value in HIV-infected
children in whom the diagnosis of intrathoracic tuberculosis is often complicated by
atypical manifestations and other HIVassociated conditions.22 In addition to
providing rapid diagnostic confirmation by
cytomorphology and/or direct visualisation
of the infecting organism, FNAB samples
may be inoculated into relevant culture that
allows more accurate organism speciation
and drug sensitivity testing.20,23
This study confirms the diagnostic value
of FNAB in routine paediatric practice. It
also demonstrates that FNAB, using a
small-gauge needle, is rapid and safe even
in resource-limited settings. Well trained
nurse aspirators can perform superficial
aspirates as an outpatient procedure, reducing the need for hospitalisation. Cytology
slides, once prepared and fixed, can be
transported easily for expert evaluation
elsewhere. FNAB is of particular value to
countries with limited resources and a high
prevalence of diseases such as tuberculosis
which can be treated successfully at primary
health care level, while identifying patients
who require referral to a regional or tertiary
care centre.
Acknowledgments
We are grateful to Dr M. Kidd, Department
of Statistics, University of Stellenbosch, for
assistance with the statistical analysis and to
45
70
C. A. Wright et al.
Mr D. Geiger for preparing the Microsoft
Excel spreadsheet.
References
1 Silverman JF, Gay M. Fine-needle aspiration and
surgical
pathology
of
infectious
lesions.
Morphologic features and the role of the clinical
microbiology laboratory for rapid diagnosis. Clin
Lab Med 1995; 15:251–78.
2 Buchino JJ, Lee HK. Specimen collection and
preparation in fine-needle aspirations in children.
Am J Clin Pathol 1998; 109 (4 suppl 1):S4–8.
3 Wakely PE, Kardos TF, Frable WJ. Application of
fine needle aspiration biopsy to paediatrics. Hum
Pathol 1988; 19:1383–6.
4 Howell LP. Changing role of fine needle aspiration
in the evaluation of paediatric masses. Diagn
Cytopathol 2001; 24:65–70.
5 van Coppenraet ESB, Lindeboom JA, Prins JM,
Peeters MF, Claas ECJ, Kuijper EJ. Real-time PCR
assay using fine-needle aspirates and tissue biopsy
specimens for rapid diagnosis of mycobacterial
lymphadenitis in children. J Clin Microbiol 2004;
42:2644–50.
6 Smith MB, Katz R, Black CT, Cangir A,
Andrassy RJ. A rational approach to the use of
fine-needle aspiration biopsy in the evaluation of
primary and recurrent neoplasms in children.
J Pediatr Surg 1993; 28:1245–7.
7 Martins MR, da Cunha Santos G. Fine needle
aspiration in the diagnosis of superficial lymphadenopathy: a 5 year Brazilian experience. Diagn
Cytopathol 2006; 34:130–4.
8 Handa U, Mohan H, Bal A. Role of fine needle
aspiration cytology in evaluation of paediatric
lymphadenopathy. Cytopathology 2003; 14:66–9.
9 DeMay RM. The Art and Science of Cytopathology.
Chicago, IL: ASCP Press, 1996.
10 Hesseling PB, Broadhead R, Molyneux E, et al.
Malawi pilot study of Burkitt’s lymphoma treatment. Med Pediatr Oncol 2003; 41:532–40.
11 Jaffe ES, Harris NL, Stein H, Vardiman JW.
Pathology and genetics of tumours of haematopoietic and lymphoid tissues. In: World Health
Organization Classification of Tumours. Lyon,
France: International Agency for Research on
Cancer, 2001; 120–53.
12 Ljung BM, Drejet A, Chiampi N, et al. Diagnostic
accuracy of fine-needle aspiration biopsy is determined by physician training in sampling technique.
Cancer 2001; 93:63–8.
13 Wright CA, Michelow P, Harnekar A, Poole J,
Leiman G. FNA in Paediatric Oncology. XV
International Academy of Cytology Congress,
Santiago, Chile, 2004.
14 Wright CA, Burgess SM, Geiger D, Wasserman E,
van den Burg M. The Diagnosis of Mycobacterial
Lymphadenitis in Children: Is Fine Needle Aspiration
the Way to Go? 46th Conference of the Federation
of South African Societies of Pathology, Durban,
South Africa, 2006.
15 Ersöz C, Polat A, Serin MS, Soylu L, Demircan O.
Fine needle aspiration (FNA) cytology in tuberculous lymphadenitis. Cytopathology 1998; 9:201–7.
16 Marais BJ, Gie RP, Schaaf HS, Donald PR, Beyers
N, Starke J. Childhood pulmonary tuberculosis—
old wisdom and new challenges. Am J Resp Crit
Care Med 2006; 173:1078–90.
17 Marais BJ, Gie RP, Schaaf HS, Hesseling AC,
Enarson DA, Beyers N. The spectrum of diseases in
children treated for tuberculosis in a highly endemic
area. Int J Tuberc Lung Dis 2006; 10:732–8.
18 Seth V, Kabra SK, Semwal OP, Mukhopadyaya S,
Jensen RL. Tubercular lymphadenitis: clinical
manifestations. Indian J Pediatr 1995; 62:565–70.
19 Marais BJ, Wright CA, Schaaf HS, et al.
Tuberculous lymphadenitis as a cause of persistent
cervical lymphadenopathy in children from a
tuberculosis-endemic area. Pediatr Infect Dis J
2006; 25:142–6.
20 Wright CA, van Zyl Y, Burgess SM, Blumberg L,
Leiman G. Autofluorescence of mycobacteria on
lymph node aspirates—a glimmer in the dark?
Diagn Cytopathol 2004; 30:257–60.
21 Marais BJ, Pienaar J, Gie RP. Kaposi sarcoma with
upper airway obstruction and bilateral chylothoraces. Pediatr Infect Dis J 2003; 22:926–8.
22 Jeena PM, Coovadia HM, Hadley LG, Wiersma R,
Grant H, Chrystal V. Lymph node biopsies in HIVinfected and non-infected children with persistent
lung disease. Int J Tuberc Lung Dis 2000; 4:139–46.
23 Silverman JF, Gay RM. Fine-needle aspiration and
surgical
pathology
of
infectious
lesions.
Morphologic features and the role of the clinical
microbiology laboratory for rapid diagnosis. Clin
Lab Med 1995; 15:251–78.
46
Fine Needle Aspiration Biopsy - A First Line Diagnostic Procedure in Pediatric
Tuberculosis suspects with Peripheral Lymphadenopathy?
Colleen A Wright, MD FRCPatha; Anneke C Hesseling, MD, PhDb ; Colleen Bamford, MD,
FCPathc; Steven Michael Burgess, PhDd; Rob Warren, PhD e; and Ben J Marais, MD, PhDb
a
Division of Anatomical Pathology, Department of Pathology, University of Stellenbosch and
NHLS Tygerberg Hospital, Tygerberg, South Africa
b
Desmond Tutu TB Center and/or the Department of Pediatrics and Child Health,
Stellenbosch University, Tygerberg, South Africa
c
Division of Medical Microbiology, Department of Pathology, University of Stellenbosch and
NHLS Tygerberg Hospital, Tygerberg, South Africa
d
Graduate School of Business, University of Cape Town, Cape Town, South Africa
e
NRF Centre of Excellence in Biomedical Tuberculosis Research / MRC Centre for Molecular
and Cellular Biology, Division of Molecular Biology and Human Genetics, University of
Stellenbosch, South Africa
SUMMARY
OBJECTIVE To evaluate the diagnostic yield and time to diagnosis of fine needle aspiration
biopsy (FNAB) versus routine respiratory specimens collected from children with a palpable
peripheral lymph node mass and symptoms suspicious of tuberculosis.
DESIGN
We performed a retrospective review of laboratory records at Tygerberg
Hospital over a 4-year period, from January 2003 to December 2006. All children (<13 years)
in whom an FNAB and other mycobacterial specimens were collected as part of their
diagnostic workup were included.
47
RESULTS
In 95 children the following specimens were collected; FNAB 95, gastric
aspirates 142, other respiratory specimens 36, non-respiratory specimens 26. Mycobacterial
disease was diagnosed in 70/95 (73.3%) patients. Children without respiratory specimens (n =
6) and/or with M bovis BCG disease (n = 15) were excluded from comparative analysis. In the
remainder, FNAB was positive in 45/74 (60.8%) versus any respiratory specimen in 29/74
(39.2%; p <0.001). The mean time to bacteriologic diagnosis with FNAB was 7.1 days (95%
CI 4.2-10.1) compared to 22.5 days (95% CI 15.8-29.1) for any respiratory specimen.
CONCLUSION
FNAB is a simple, rapid and effective modality to achieve
confirmation of mycobacterial disease in pediatric tuberculosis suspects with a palpable
peripheral lymph node mass.
INTRODUCTION
Pediatric tuberculosis (TB) contributes significantly to the global TB disease burden.1-2
Children suffer severe TB-related morbidity and mortality in endemic areas, but contribute
little to disease transmission and the maintenance of the TB epidemic.1 Consequently,
treatment of children has been a low priority in global TB control efforts, but the World
Health Organization (WHO) produced guidance for the management of childhood TB in 2006
calling for the evaluation of new techniques to improve the diagnosis.3 The Global Drug Fund
made child-friendly TB treatment formulations available to poor countries since 2008.4
The accurate diagnosis of pediatric TB remains a significant challenge, due to the low
specificity of signs and symptoms, especially in human immunodeficiency virus (HIV)infected children, the difficulty of obtaining bacteriologic specimens and the paucibacillary
nature of pediatric disease.5 New diagnostic tests for TB have been developed, such as T cell
assays and nucleic amplification tests, but these are all poorly validated in children. 6 The
48
ability to distinguish latent infection from active disease remains problematic, particularly in
endemic areas with heavy disease burdens.
An accurate bacteriologic diagnosis in children with TB is needed to initiate effective
therapy and has become even more important with the transmission of multidrug-resistant
(MDR) and extensively drug-resistant (XDR) TB to children.7-8 Obtaining bacteriologic
specimens from children is difficult and diagnostic yield is usually low. Sputum smear
microscopy is positive in less than 10-15% of children with probable pulmonary TB, with
reported culture yields between 30-40%. 9-11 Various methods have been described to obtain
respiratory specimens from children who are unable to expectorate, 12-14 with no international
consensus on what is best.
TB lymphadenitis is the most common extrapulmonary manifestation of TB in
children from endemic areas,15 accounting for almost 50% of extrapulmonary disease.16 Fine
needle aspiration biopsy (FNAB) is the diagnostic modality of choice in children with
persistently enlarged cervical nodes in TB endemic areas.17 FNAB is increasingly utilized in
paediatric oncology patients, 18-19 and multiple studies have reported the value of FNAB for
TB diagnosis in adults, 20-22 but it remains underutilized as a diagnostic tool in TB suspects.
The aims of this study were to compare the diagnostic yield and time to diagnosis of FNAB
compared to routine respiratory specimens collected in child TB suspects with a palpable
peripheral lymph node mass.
METHODS
We performed a retrospective review of laboratory records at Tygerberg Hospital, Cape Town
South Africa, from January 2003 to January 2007. Only children less than 13 years of age
(eligible for access to pediatric services) were included. Since this was a laboratory-based
study we were unable to review clinical information and all diagnostic tests were requested at
49
the discretion of the attending clinician. Standard practice at our institution is to refer
children with a persistent peripheral lymph node mass (>1-2x1-2cm) for FNAB, while
respiratory specimens are collected in most pediatric TB suspects.
Children who underwent FNAB were identified using laboratory records of the
National Health Laboratory Services. Those in whom another specimen for mycobacterial
culture was obtained within a 30 day time window either side of the FNAB collection date
were included in the analysis. Cases with Mycobacterium bovis bacille Calmette-Guerin
(BCG) disease were excluded from the comparative analysis. These children typically present
with right axillary adenitis ipsilateral to the vaccination site, which is nearly pathognomonic.23
Inclusion of these children would have biased results in favour of FNAB.
FNAB was done in the ward or as an outpatient procedure by a trained pathologist
following standard protocol.24 The lymph node mass was stabilized by the pathologist
performing the aspiration biopsy using a 23g or 25g needle attached to a 10ml syringe,
applying constant suction of no more than 2ml. Two smears were prepared from each
aspirate: one fixed with commercial cytology fixative for Papanicolaou staining and the other
air-dried for Giemsa and ZN staining. Thereafter the needle and syringe was rinsed by
withdrawing an aliquot of liquid growth media into the syringe and discharging the contents
back into a mycobacterial growth indicator tube (MGIT, Beckton Dickinson, USA). The
inoculated MGIT tubes were transported to the microbiology laboratory within 2 hours, where
standard PANTA supplement was added and the tubes incubated for 42 days. Positive
cultures were identified as M. tuberculosis complex by polymerase chain reaction (PCR).25
In the cytology laboratory, smears were evaluated for adequacy, defined as sufficient
cells and/or necrotic material to allow a definitive diagnosis. A modified ZN stain was used
on the Giemsa stained smears to detect acid fast bacilli.24 Fluorescence microscopy was
carried out on Papanicolaou stained smears using a Zeiss Axiophot microscope with a
50
fluorescent attachment and a wide-band blue excitation filter (450–480 nm). Papanicolaou
stained mycobacteria fluoresce as brilliant yellow bacilli, thin and slightly curved with polar
enhancement and a length of 2.0 –2.7 microns.26 M. tuberculosis, M. bovis, M. bovis BCG and
non-tuberculous mycobacteria (NTM) are morphologically indistinguishable, but in this TBendemic setting M. bovis is hardly ever detected on culture, M. bovis BCG usually has a
clinical pathognomonic presentation, 27 and disease caused by NTM is relatively rare.
Early morning gastric aspirates were obtained following overnight admission to the
pediatric ward. A nasogastric tube was passed into the stomach and the contents aspirated; 510 ml of normal saline inserted if required until a total volume of at least 5ml was aspirated.
Gastric aspirates were placed in a sterile tube and neutralized with an equal volume of sodium
carbonate. Induced sputa and/or nasopharyngeal aspirates were collected by a trained
physiotherapist, but this was rarely done. Sputum induction was undertaken after 3 hours of
fasting. Children were pre-treated with 200ug of salbutamol via metered dose inhaler and
spacer. A jet nebulizer attached to oxygen at a flow rate of 5 l/min delivered 5 ml of 5%
sterile saline for 15 minutes. Thereafter chest physiotherapy was applied before obtaining
sputum by expectoration or by nasopharyngeal suctioning, using a mucus extractor.
Oropharyngeal suction was used if the child had excessive nasal secretions.
All microbiology specimens were submitted to the laboratory within 6 hours. Smear
microscopy, using either Ziehl-Neelsen or auramine stains, was only performed on
expectorated sputum samples. Following N-acetyl-L- cysteine - sodium hydroxide
decontamination, sputum specimens were inoculated into MGIT and incubated for 42 days.
Further procedures for identification of mycobacteria were as described for the FNAB
samples. All procedures were carried out according to standard operating procedures in a
Biosafety Level 2 laboratory.
51
The reference standard for disease on FNAB was cytomorphology consistent with
mycobacterial infection plus mycobacteria visualized with ZN stain and/or autofluorescence,
and/or if M. tuberculosis was identified on culture. Given the low rates of NTM in the study
population culture negative cases with cytomorphological and mycobacterial proof of disease
were included in the TB group. For respiratory specimens the reference standard was acid fast
bacilli (AFB) detected on sputum smear microscopy or identification of M. tuberculosis by
culture in at least one specimen. For comparative purposes multiple respiratory specimens
were regarded as a single test procedure, since we aimed to compare the value of routine
respiratory specimen collection as a diagnostic procedure to FNAB.
We assessed whether the comparative bacteriological yield of FNAB (cytology and/or
culture) was higher than other methods using Pearson’s chi-square (X2) and Fisher’s exact
tests 28. All analyses were conducted using Statistica Version 8.29 Ethics approval was
obtained from the Institutional Review Board of Stellenbosch University (N08/09/240).
RESULTS
FNAB and additional mycobacterial specimens were collected from 95 children, ranging in
age from 2 months to 11 years. Figure 1 provides a flow diagram of patients and specimens
included in the study. There was no significant difference in gender distribution (49 male vs.
46 females) and 48% of children were less than one year of age. The HIV-infection status was
unknown in 28 (29.5%) children while 33 (34.7%) were HIV-infected and 34 (35.8%) were
HIV-uninfected. Mycobacteria were identified in 70 (73.7%) children. Table 1 reflects the
biopsy site, bacteriologic yield and mycobacterial species identified on FNAB; no NTM were
identified.
Table 2 reflects the bacteriologic yield and mycobacterial species identified for all
specimen types; of the 57 specimens speciated 41 (71.9%) were M. tuberculosis, 15 (26.3%)
52
M. bovis BCG and 1 (1.8%) NTM. A total of 143 gastric aspirates were collected from 80
(84.2%) children, with only 11 children (13.8%) submitting the desired 3 specimens.
Additional respiratory specimens were collected in 15 patients; of whom 7 (mean age 9.3
years) provided expectorated sputum. Non-respiratory specimens, including cerebrospinal
fluid, pus swabs and blood, were collected in 21 patients (26 specimens). All non-respiratory
specimens, including 6 patients in whom no respiratory specimens were collected, and M
bovis BCG cases were excluded from the comparative analysis. M bovis BCG was isolated in
15 children; in 2 cases the site of aspiration was soft tissue rather than an axillary lymph node.
One third of these children (5 cases) had distant or disseminated BCG disease as evidenced by
the isolation of M. bovis BCG from gastric and/or nasopharyngeal aspirates. All were
immunocompromised: 4 were HIV-infected and the other had Severe Combined Immune
Deficiency.
Of the 45 patients diagnosed with TB using the defined FNAB reference standard, 36
(80.0%) were confirmed to be M. tuberculosis on culture. Of the 30 diagnosed from
respiratory specimens, 29 (96.7%) were confirmed to be M. tuberculosis on culture and 1
adolescent child was sputum smear (AFB) positive. For comparative purposes the collection
of multiple respiratory specimens was regarded as a single test procedure. The bacteriological
yield of FNAB compared to any combination of respiratory specimens is reported in Table 3;
being higher compared to gastric aspirates (p < 0.001) and all respiratory specimens combined
(p < 0.001).
We also assessed differences in the mean time to diagnosis for FNAB versus the other
specimen types (where time to diagnosis is time to bacteriologic confirmation from onset of
the test procedure). Whether the timing of additional mycobacterial cultures occurred before,
after or simultaneously with FNAB collection did not influence the significance of the yield
or time to diagnosis comparisons. Collecting multiple respiratory specimens was regarded as
53
a single test procedure. Accepting an average time to microscopy diagnosis of 2 days, FNAB
(7.1 days, 95% confidence interval = 4.2 - 10.1 days) requires less than one-third the time of
respiratory cultures (22.5 days, 95% CI = 15.8 – 29.1 days). This difference is illustrated in
Figure 2. The one outlier is the child whose TB was confirmed by sputum smear-microscopy.
The mean time to diagnosis for gastric aspirate culture (the most commonly performed test)
was 21.4 days (95% CI = 15.2 – 27.6 days).
DISCUSSION
In the current study FNAB provided a better yield in child TB suspects with palpable
peripheral lymph node masses than the collection of multiple respiratory specimens.
Children referred with suspected mycobacterial lymphadenitis routinely have a tuberculin
skin test (TST) performed and return to the clinic or hospital for this to be read within 48–72
hours. If FNAB is performed with the initial visit, microscopy results should be available at
the time of the return visit 48-72 hours later, enabling the commencement of TB therapy
pending the results of culture and speciation.
The true extent of the pediatric TB disease burden is poorly documented, since more
than 80% of children with TB are sputum smear-negative and direct smear microscopy is
often the only diagnostic modality available in countries with limited resources. The
diagnostic dilemma is further compounded by the HIV pandemic, 27 HIV-associated disease
and TB may have common clinical and radiological presentations, in addition tests for TB
infection such as the TST and newer T-cell assays have poor sensitivity in
immunocompromised children.2, 6, 10 Two to three fasting gastric aspirates or a single induced
sputum specimen have traditionally been advised for bacteriological confirmation of TB.12
However, the yields observed in everyday practice deviate considerably from those achieved
under trial conditions, which emphasizes the value of “real-life” retrospective analysis.
54
The low number of patients in whom gastric aspirates were collected on 3 consecutive
days reflects practical difficulties. Severe limitations of in-patient beds and staff and the
reluctance of parents to remain in hospital with their children for more than 1-2 days, since
they often have other children at home to care for. In 5 children induced sputa and/or
nasopharyngeal aspirates were collected as part of a separate specimen collection study,
comparing the yield of gastric aspirates, nasopharyngeal aspirates and induced sputum. We
included these specimens to provide a comprehensive overview of all the respiratory
specimens collected. Inclusion of these specimens may have introduced bias, but numbers
were small and did not influence the outcome. If anything these additional specimens would
have favored the bacteriological yield of routine respiratory specimens. Development of
improved respiratory specimen collection techniques that can be performed on an outpatient
basis require further exploration. However our study demonstrates the value of FNAB as an
outpatient diagnostic procedure in TB suspects with a palpable peripheral lymph node mass.
Enlarged peripheral lymph nodes provide a valuable opportunity to establish a
definitive diagnosis of TB in HIV-infected children.30 The majority of HIV-infected children
with persistent lung disease have concurrent persistent peripheral lymphadenopathy, 30 which
provides a neglected sample collection opportunity in this problematic group. In a recent
South African study, 22 of 95 FNAB performed in mass lesions in HIV–infected children
were positive for mycobacterial infection.31 FNAB has been established as an easy and
reliable outpatient procedure for the diagnosis of palpable superficial masses in children, and
is ideally suited for use in resource limited settings. It is simple, inexpensive, requires limited
infrastructure and can be taught to medical and nursing staff.32 More widespread utilization is
limited by a lack of experienced pathologists, but slides are easily transportable to regional or
even international centers for diagnosis. 33
55
If the correct technique is followed, it is associated with minimal complications.33
There were no adverse events or complications recorded with any of the FNAB’s performed
during this study. In a previous study we demonstrated that using FNAB with
cytomorphology, autofluorescence and culture, provided a rapid and definitive diagnosis in
the vast majority of children.24 Culture alone produced a significantly lower yield, since it is
influenced by factors such as inefficient sampling, contamination, variable bacterial load, and
prior commencement of TB therapy. Rapid microscopy confirmation was achieved in 34/45
(75.5%) patients ultimately diagnosed with M tuberculosis using a combination of
microscopy and/or culture positivity. Previous FNAB studies from this TB endemic setting, as
well as the current study, demonstrated no NTM disease, 17, 24, 26, 31 while M bovis BCG was
only cultured in children less than 2 years age with a clinically pathognomonic presentation.
Although rapid microscopy diagnosis allows a fairly accurate diagnosis and permits
immediate commencement of therapy, culture is advisable as this enables speciation and drug
susceptibility testing.
Compared to conventional respiratory specimens FNAB provided superior diagnostic
yields and a significant reduction in time to TB diagnosis. Rapid bacteriological diagnosis
has important benefits for patient management, also in settings not endemic for TB where
NTM adenitis may be more common.34-35 This study reviewed laboratory records only,
therefore, we are unable to provide detailed clinical descriptions or explore potential
differences in patient characteristics that may have influenced the sequencing and/or yield of
various diagnostic tests. However, it provides an assessment of every day (“real life”)
practice without any intentional bias. Despite these limitations, we believe FNAB should be
regarded as a first line diagnostic modality in child TB suspects with a palpable peripheral
lymph node mass.
56
ACKNOWLEDGEMENTS
The authors would like to thank Professor Martin Kidd (Stellenbosch University) for
statistical analysis. This study is in partial fulfillment of a PhD thesis.
57
REFERENCES
[1]
Marais BJ, Gie RP, Schaaf HS, et al. Childhood pulmonary tuberculosis: old wisdom
and new challenges. Am J Respir Crit Care Med 2006;173:1078.
[2]
Cotton MF, Schaaf HS, Hesseling AC, Madhi SA. HIV and childhood tuberculosis:
the way forward. Int J Tuberc Lung Dis 2004;8:675-682.
[3]
World Health Organization. A research agenda for childhood tuberculosis. Geneva:
World Health Organization; 2007. Report No.: WHO/HTM/TB/2007.381.
[4]
Gie CP, Matiru RH. Supplying quality-assured child-friendly anti-tuberculous drugs to
children. Int J Tuberc Lung Dis 2009;13:277-278.
[5]
Eamranond P, Jaramillo E. Tuberculosis in children: reassessing the need for
improved diagnosis in global control strategies. Int J Tuberc Lung Dis 2001;5:594-603.
[6]
Marais BJ, Pai M. New approaches and emerging technologies in the diagnosis of
childhood tuberculosis. Paediatr Resp Rev 2007;8:124-133.
[7]
Schaaf HS, Marais BJ, Whitelaw A, et al. Culture-confirmed childhood tuberculosis in
Cape Town, South Africa: a review of 596 cases. BMC Infectious Diseases 2007:140.
[8]
Chaisson RE, Martinson NA. Tuberculosis in Africa - combating an HIV-driven crisis.
N Engl J Med 2008;358:1089-1092.
[9]
Newton SM, Brent AJ, Anderson S, Whittaker E, Kampmann B. Paediatric
tuberculosis. Lancet Infect Dis 2008;8:499-510.
[10]
Shingadia D, Novelli V. Diagnosis and treatment of tuberculosis in children. Lancet
Infect Dis 2003;3:624-632.
[11]
Marais BJ, Graham SM, Cotton MF, Beyers N. Diagnostic and management
challenges for childhood tuberculosis in the era of HIV. J Infect Dis 2007;196:76-85.
[12]
Coulter JBS. Diagnosis of pulmonary tuberculosis in young children. Ann Trop
Paediatr 2008;28:3-12.
58
[13]
Owens S, Abdel-Rahmen IE, Balyejusa S, et al. Nasopharyngeal aspiration for
diagnosis of pulmonary tuberculosis. Arch Dis Child 2007;92:693-696.
[14]
Zar HJ, Hanslo D, Appoles P. Induced sputum versus gastric lavage for
microbiological confirmation of pulmonary tuberculosis in infants and young children: a
prospective study. . Lancet 2005;365:130-134.
[15]
Marais BJ, Gie RP, Schaaf HS, et al. The spectrum of childhood tuberculosis in a
highly endemic area. Int J Tuberc Lung Dis 2006;10:732-738.
[16]
Cruz AT, Starke JR. Clinical manifestations of tuberculosis in children. Paediatr
Respir Rev 2007;8:107-117.
[17]
Marais BJ, Wright CA, Schaaf HS, et al. Tuberculous Lymphadenitis as a Cause of
Persistent Cervical Lymphadenopathy in Children From a Tuberculosis-Endemic Area.
Pediatr Infect Dis J 2006;25:142-146.
[18]
Drut R, Drut RM, Pollono D, et al. Fine-Needle Aspiration Biopsy in Paediatric
Oncology Patients. A Review of Experience with 829 Patients. Paediatr Haematol Oncol
2005;27:370-376.
[19]
Wakely PE, Kardos TF, Frable WJ. Application of fine needle aspiration biopsy to
pediatrics. Hum Pathol 1988;19:1383-1386.
[20]
Nayak S, Mani R, Kavatkar AN, Puranik SC, Holla VV. Fine-needle aspiration
cytology in lymphadenopathy of HIV-positive patients. Diagn Cytopathol 2003;29:146-148.
[21]
Kocjan G, Miller R. The cytology of HIV-induced immunosuppression. Changing
pattern of disease in the era of highly active antiretroviral therapy. Cytopathology
2001;12:281-296.
[22]
Gupta AK, Nayar M, Chandra M. Critical appraisal of fine needle aspiration cytology
in tuberculous lymphadenopathy. Acta Cytol 1992;36:391-394.
59
[23]
Hesseling AC, Rabie H, Marais BJ, et al. Bacille Calmette-Guérin Vaccine–Induced
Disease in HIV-Infected and HIV-Uninfected Children. Clin Infect Dis 2006:548–558.
[24]
Wright CA, van der Burg M, Geiger D, et al. Diagnosing Mycobacterial lymphadenitis
in children using Fine Needle Aspiration Biopsy: Cytomorphology, ZN staining and
Autofluorescence – making more of less. Diagn Cytopathol 2008;36:245-251.
[25]
De Wit D, Steyn L, Shoemaker S, Sogin M. Direct detection of Mycobacterium
tuberculosis in clinical specimens by DNA amplification. J Clin Microbiol 1990;28:24372441.
[26]
Wright CA, van Zyl Y, Burgess SM, Blumberg L, Leiman G. Auto fluorescence of
Mycobacteria on Lymph Node Aspirates – A glimmer in the dark? Diagn Cytopathol
2004;30:257-260.
[27]
Hesseling AC, Cotton MF, Jennings T, et al. High Incidence of Tuberculosis among
HIV-Infected Infants: Evidence from a South African Population-Based Study Highlights the
Need for Improved Tuberculosis Control Strategies. Clin Infect Dis 2009;48:108-114.
[28]
Agriesti A. Categorical Data Analysis. 2nd ed. Hoboken, New Jersey: Wiley 2002.
[29]
Statistica. Version 8 ed. Tulsa, Oklahoma: Statsoft, Inc.
[30]
Jeena PM, Coovadia HM, Hadley LG, et al. Lymph node biopsies in HIV infected and
non infected children with persistent lung disease. Int J Tuberc Lung Dis 2000:139-146.
[31]
Michelow P, Meyers T, Dubb M, Wright C. The utility of fine needle aspiration in
HIV positive children. Cytopathology 2008;19:86-93.
[32]
Leiman G. The Registered Nurse as Aspirator:Experience in a Busy Cytopathology
Unit. Acta Cytol 1995;39:1010.
[33]
Wright CA, Pienaar JP, Marais BJ. Fine needle aspiration biopsy: diagnostic utility in
resource-limited settings. Ann Trop Paediatr 2008; 28:65-70.
60
[34]
Hazra R, Robson CD, Perez-Atayde AR, Husson RN. Lymphadenitis Due to
Nontuberculous Mycobacteria in Children: Presentation and Response to Therapy. Clin Infect
Dis 1999;28:123-129.
[35]
Benson-Mitchell R, Buchanan G. Cervical lymphadenopathy secondary to atypical
mycobacteria in children. J Laryngol Otol 2007;110:48-51.
61
Table 1
Mycobacteria isolated on culture and site of fine needle aspiration biopsy (FNAB)
Mycobacteria
isolated a
Site of FNAB
Number
(%)
Axillary node
R
L
Not
stated
SubCervical Salivary Soft
mandibular
node
gland tissue
node
Site
not
stated
All
51/95
(54)
11
5
6
8
31
4
5
25
M.
Tuberculosis
36/51
(71)
5
5
2
8
31
4
3
22
M. bovis
BCG b
15/51
(29)
6
0
4
0
0
0
2
3
a
b
No M. bovis or Non Tuberculous Mycobacteria (NTM) were isolated
13 (87%) of children <1 year of age
62
Table 2
Specimen type, bacteriological yield and organisms cultured in 95 patients
Specimen type
Consecutive
specimens
No (%)
Total
number
1st
2nd
3rd
0
0
Culture
positive
No (%)
M. tba
BCGb
NTMc
No (%)
No (%)
No (%)
51 (54)
36 (71)
15 (30)
0
FNAB
95
95
Gastric
aspirates
143
80
52(65) 11(14)
39 (27)
30 (77)
8 (21)
1 (3)
Induced sputum
15
10
5(50)
0
6 (40)
6 (100)
0
0
Expectorated
sputum
11
7
3(43)
1(14)
5 (45)
5 (100)
0
0
Nasopharyngeal
aspirates
10
7
3(43)
0
5 (50)
4 (80)
1 (20)
0
Non–
respiratory
26d
26
0
0
3 (12)
2(67)e
1 (33)f
0
a
M.tb – Mycobacterium tuberculosis;
b
BCG – Mycobacterium bovis bacilli Calmette-Guerin;
c
NTM – Non-tuberculous mycobacteria
d
21 patients, pus swabs, urine, CSF, blood cultures;
e
Cerebrospinal fluid;
f
Pus swab.
63
Table 3
Bacteriologic yield of fine needle aspiration biopsy (FNAB) compared to routine
respiratory specimens (M bovis BCG excluded)
Routine respiratory specimens
FNABa
negative
FNABa
positive
Total
All gastric aspirates negative
25
19
44
Any gastric aspirate positive
3
21
24
Total
28
40
68
All respiratory specimensb negative
26
19
45
Any respiratory specimen positive
3
26
29
Total
29
45
74
a
Cytology and/or culture
b
Includes gastric and nasopharyngeal aspirates, as well as sputum and induced sputum
specimens
64
Figure 1
Breakdown of patients and specimens included in the analysis
95 patients
with FNAB*
and another
specimen
143 gastric aspirates
15 patients with
M. bovis BCG
36 other respiratory
specimens
Excluded
6 patients,
non-respiratory
specimens only
26 non-respiratory
specimens (excluded)
74 patients
with FNAB*
and another
respiratory specimen
68 patients with
gastric aspirates
12 patients with
other respiratory
specimens
*FNAB – Fine needle aspiration biopsy
74 patients with
FNAB*
65
Figure 2
Time to diagnosis Fine Needle Aspiration Biopsy (FNAB) versus All respiratory
specimens
Y-axis – specify units “Days”
X-axis – only “FNAB” & “All respiratory”
66
CHAPTER 5
OPTIMISING CYTOPATHOLOGICAL DIAGNOSIS
Mycobacterial Autofluorescence in Papanicolaou-Stained Lymph Node Aspirates: a
Glimmer in the Dark?
Colleen A Wright, Yvonne van Zyl, Steven M Burgess, Lucille Blumberg, Gladwyn Leiman.
Diagnostic Cytopathology. 2004;30:257-60.
Diagnosing Mycobacterial Lymphadenitis in Children Using Fine Needle Aspiration
Biopsy: Cytomorphology, ZN Staining and Autofluorescence-making more of less.
Colleen A Wright, Steven M Burgess, Mirjam van der Burg, D Geiger, Jeroen G Noordzij, Ben
J Marais
Diagnostic Cytopathology: 2008; 36: 245-251
67
Mycobacterial Autofluorescence
in Papanicolaou-Stained Lymph
Node Aspirates:
A Glimmer in the Dark?
Colleen A. Wright, FRCPath,1* Yvonne van Zyl, C.T.,1
Steven M. Burgess, Ph.D.,2 Lucille Blumberg, M.D.,3
and Gladwyn Leiman, FRCPath1
This study was undertaken to determine the value of incorporating
fluorescence into cytopathological evaluation of lymph node fineneedle aspiration (FNA) specimens suspected of harboring mycobacterial species. The study population consisted of 1,044 HIVpositive and -negative patients referred for FNA to the
cytopathology unit of a South African medical school located in a
very high HIV prevalence region. Each aspirate was assessed on
routine Papanicolaou-stained slides for morphologic characteristics of mycobacterial infection. The same glass slides were then
viewed under fluorescent microscopy to determine the presence or
absence of mycobacterial autofluorescence. Using multivariate
analysis, results of both cytology and fluorescence were compared
with mycobacterial culture as the final arbiter of the presence of
organisms. In this large clinical study, compared with culture,
cytomorphology showed sensitivity of 84.9%, but low specificity of
only 50.9%. Fluorescence demonstrated lower sensitivity of
65.9%, but improved specificity of 73.0%. Taken together, positivity of both cytology and fluorescence improved specificity to
81.8%. Fluorescent microscopy is rapid, inexpensive, and cost-
1
Cytopathology Unit, School of Pathology of the South African Institute
for Medical Research and University of the Witwatersrand, Johannesburg,
South Africa
2
Department of Commerce, University of the Witwatersrand, Johannesburg, South Africa
3
Department of Microbiology, School of Pathology, University of the
Witwatersrand, Johannesburg, South Africa
Colleen A. Wright’s current address is Department of Anatomical Pathology, Stellenbosch University, Cape Town, South Africa.
Steven M. Burgess’s current address is Graduate School of Business,
University of Cape Town, Cape Town, South Africa.
Lucille Blumberg’s current address is National Institute for Communicable Diseases, National Health Laboratory Services, Johannesburg, South
Africa
Gladwyn Leiman’s current address is Department of Pathology, University of Vermont, Burlington, Vermont.
The South African Institute for Medical Research is now the National
Health Laboratory Service of South Africa.
*Correspondence to: Colleen A. Wright, Department of Anatomical
Pathology, University of Stellenbosch, P.O. Box 19063, Tygerberg 7505,
South Africa. E-mail: cawr@gerga.sun.ac.za
Received 30 May 2003; Accepted 29 September 2003
DOI 10.1002/dc.20009
Published online in Wiley InterScience (www.interscience.wiley.com).
© 2004
WILEY-LISS, INC.
effective; neither radioactive materials nor further staining are
required. It is felt that this methodology would be of diagnostic
benefit if used on morphologically suspicious samples in areas
with a high prevalence of HIV and mycobacterial infections.
Appropriate therapy could be commenced within hours of FNA,
with reduction in the current number of patients lost to follow-up
while awaiting results of culture. The technique is readily extended
to other FNA types such as deep organ aspirates. Autofluorescence
of organisms specifically requires usage of Papanicolaou staining;
the technique cannot be used in histopathologic specimens stained
with hematoxylin-eosin. Diagn. Cytopathol. 2004;30:257–260.
©
2004 Wiley-Liss, Inc.
Key Words: mycobacteria; tuberculosis; fluorescence; cytodiagnosis; HIV
The human immunodeficiency virus (HIV) epidemic in
South Africa has attracted considerable attention. Prevalence figures in the general population are unknown, but
24.5% of women attending antenatal clinics nationwide are
HIV-positive; the corresponding figure for the province in
which this study was undertaken is 29.4% (Department of
Health National HIV and Syphilis seroprevalence survey of
women attending public antenatal clinics in South Africa,
2000; www.doh.gov.za/docs/reports/2000/hivreport.html).
It is estimated that more than 60% of medical admissions in
academic hospitals are due to opportunistic infections in
patients with acquired immunodeficiency disease (AIDS).
The most common of these opportunistic infections in South
African adults with AIDS is Mycobacterium tuberculosis
(MTB). In the year 2000, 150,696 new MTB cases were
reported to the national registry, including 31,058 extrapulmonary infections (personal communication, Tuberculosis
Directorate, SA National TB Control Program). By 2001,
MTB incidence in the general population was approximately 300 per 100,000; at least 50% of new MTB cases
were in retroviral-positive patients (personal communication, SA National Department of Health). Dual HIV and
Diagnostic Cytopathology, Vol 30, No 4
257
68
WRIGHT ET AL.
MTB epidemics place overwhelming demands on health
care resources in any developing country. Expected life
span is reduced by AIDS (down from 68 years to 48 years
in South Africa between 1990 and 2000).1,2 By dominating
hospital admissions, the HIV/MTB population limits investigation and treatment of non-HIV patients with potentially
curable disease entities compatible with normal longevity.
MTB diagnosis may be challenging, even in high prevalence areas; it is modified by HIV status and compounded
by atypical symptoms, overlap in presentation of other
opportunistic infections, frequency of extrapulmonary involvement, and rapid progression.3
Traditionally, laboratory diagnosis of mycobacterial infection has been the purview of the microbiologist; it has relied on
positive culture or microscopic recognition of the organism by
Ziehl-Nielsen staining.4 Fluorescence microscopy using auramine-rhodamine was described as early as 1937 and generally
has been shown to be superior to Ziehl-Nielsen staining,5–7 but
involves the use of toxic and carcinogenic substances. A substantial benefit would derive from any new, inexpensive, exposure-free diagnostic procedure, enabling early commencement of anti-MTB treatment. Mycobacteria (as well as
Pneumocystis carinii, fungi, and certain bacteria) show brilliant autofluorescence in Papanicolaou-stained smears.8 –10 The
method is quick and inexpensive and can be focussed on those
specimens felt to harbor the above infections on morphologic
grounds. This suggests that fluorescence microscopy of Papanicolaou-stained smears could provide a rapid, safe, and inexpensive technique for confirmation of mycobacterial infection
in cytological specimens. The present study explores this opportunity by investigating the ability of cytomorphology and
fluorescence microscopy of Papanicolaou-stained smears from
superficial lymph nodes obtained by fine-needle aspiration
(FNA), individually and together, to identify mycobacterial
infections.
Materials and Methods
A prospective study was performed in the teaching hospitals of
the University of the Witwatersrand in Johannesburg, South
Africa, during the period January 1998 to December 2001. The
sample consisted of 1,044 superficial lymph node aspirates
from patients with clinically suspected mycobacterial lymphadenopathy. In this center, retroviral serology is not performed
without patient consent; serology was thus not available in
many cases. In the 330 cases in which retroviral status was
known, 318 (96%) were positive; a high proportion of the
remaining patients were clinically suspected to be retroviruspositive. Patients were referred to FNA clinics operated by the
cytopathology unit at four teaching hospitals. Procedures were
performed using 22 gauge needles attached to 10 ml syringes.
At least two needle passes were performed on each node
aspirated. Direct smears were made and were either air-dried
for Diff-Quik staining or fixed with commercial spray fixative
258
Diagnostic Cytopathology, Vol 30, No 4
Fig. C-1. Dark-field illumination of characteristic MTB organism demonstrating autofluorescence in a lymph node aspirate (Papanicolaou, ⫻1,000).
for Papanicolaou staining. Onsite rapid staining was not undertaken. The needle was rinsed in BACTEC 12B medium
(Becton-Dickinson, Mountain View, CA), a commercially
manufactured mycobacterial culture system. Slides were returned to the central cytopathology laboratory for same-day
processing.
The air-dried Diff-Quik-stained and alcohol-fixed Papanicolaou-stained smears were screened and evaluated for
adequacy and content. All neoplastic samples were excluded from the study, as were inadequate slides and those
showing benign nonreactive, noninflammatory morphology.
The remainder were reported as cytologically suggestive
(positive) or nonsuggestive (negative) of mycobacterial
lymphadenitis. This cytomorphologic evaluation was made
on previously established criteria: amorphous granular necrotic debris, epithelioid histiocytes, and multinucleated
Langhans-type giant histiocytes, in a background population of reactive lymphocytes and plasma cells. In the presence of AIDS, the cytological picture is more frequently that
of necrosis only, or necrosis associated with an acute inflammatory infiltration, without attendant lymphocytic or
histiocytic components.
Without any additional processing, the Papanicolaoustained smears were then screened by one author (Y.v.Z.)
using an Olympus BX40 microscope with a fluorescent
attachment and the following filters: wide-band blue excitation filter (450 – 480 nm), dichromatic splitter (500 nm),
and barrier filter (515 nm). Slides were viewed at ⫻400
magnification. This displayed the MTB organisms as brilliant yellow fluorescent rods, which were thin and slightly
curved with light ends and a uniform length of 2.0 –2.7
microns (Fig. C-1). Mycobacterium tuberculosis is differentiated from Mycobacterium avium, which is shorter (1.2–
1.8 microns) and variably fluorescent.
69
MYCOBACTERIAL AUTOFLUORESCENCE
Table I. Comparative Results of MTB Detection by Cytodiagnosis and
by Culture*
Culture
Cytodiagnosis
Negative
Positive
Total
Negative
Positive
Total
228 (22%)
220 (21%)
448 (43%)
90 (8%)
506 (49%)
596 (57%)
318 (31%)
726 (70%)
1,044 (100%)
*Sensitivity, 84.9%; specificity, 50.9%; negative predictive value, 71.7%;
positive predictive value, 69.7%; accuracy, 69.3%.
Table II. Comparative Results of Detection of MTB by
Autofluorescence and by Culture*
Culture
Fluorescence
Negative
Positive
Total
Negative
Positive
Total
327 (31%)
121 (12%)
448 (43%)
203 (19%)
393 (38%)
596 (57%)
530 (50%)
514 (50%)
1,044 (100%)
*Sensitivity, 65.9%; specificity, 73.0%; negative predictive value, 67.7%;
positive predictive value, 76.5%; accuracy, 69.3%.
Table III. Summary of Statistical Analyses for Detection of MTB by
Cytodiagnosis and by Autofluorescence, Separately and Together (%)
Sensitivity
Specificity
Negative predictive value
Positive predictive value
Accuracy
Cytodiagnosis
Autofluorescence
Both
84.9
50.9
71.7
69.7
69.3
65.9
73.0
61.7
76.5
69.3
56.7
81.8
58.7
80.5
69.4
Results
Retroviral status (where known) and the diagnostic outcome
of each of the three tests (i.e., cytomorphology, fluorescence, and culture) were code-captured in a standard portable computer spreadsheet using a standard indicator coding
scheme (1 ⫽ positive; 0 ⫽ negative). The laboratory’s
unique case reference number was also captured for quality
control purposes. Detailed results appear in Tables I–III.
The results indicate a strong and highly significant association between cytology and culture diagnoses (Cramer’s
V ⫽ 0.59; P ⬍ 0.001) and a virtually equivalent association
between fluorescence and culture (Cramer’s V ⫽ 0.57; P ⬍
0.001). Thus, the overall accuracy of the techniques, separately and together, is virtually identical at almost 70%.
Cytology and fluorescence offer differing benefits as regards sensitivity and specificity. Cytology is suggestive of
84.9% of culture-positive cases, yields few false negatives,
but has low specificity of 50.9%. Fluorescence has lower
sensitivity of 65.9%, yields few false positive readings, and
thereby raises specificity to 73.0%. Taken together, cytology and autofluorescence refine the diagnostic group, incorporating lower sensitivity at 56.7%, but significantly improving specificity to 81.8%. Clinical implications are that
three of five TB patients undergoing FNA could be placed
on appropriate and immediate antituberculous therapy. At a
later stage of patients so medicated, one of four would not
be confirmed on traditional culture as harboring active mycobacterial infection.
Discussion
The diagnosis of tuberculosis by cytomorphology is not
new. It is a necrotizing granulomatous infection, which
cytologically demonstrates the microscopic equivalent of
caseous necrosis, i.e., a granular-appearing necrotic background, together with mature lymphocytes, tangles of epithelioid histiocytes, and giant multinucleated Langhanstype histiocytes. This cytological appearance, however, is
neither sensitive nor specific for MTB. In recent years, the
advent of HIV-associated tuberculosis has altered the characteristic cytologic picture, mainly by paucity or even absence of cellular response, i.e., fewer epithelioid and
multinucleated histiocytes. Recent literature has shown that
FNAs, particularly from large nodes with necrotic centers,
may contain no cellular response at all, i.e., no epithelioid or
multinucleated histiocytes, and may even lack lymphocytes
if the entire node has undergone necrosis.11,12 Aspirates may
then be entirely necrotic or may contain debris together with
a secondary influx of neutrophils. The cytology then resembles an abscess more than it does classic tuberculosis. However, in high-incidence regions, sufficient experience has
now been gained with new and varied presentations of
tuberculosis in lymph node aspirates to raise the cytological
suspicion of MTB in appropriate clinical settings. ZiehlNielsen staining may be performed on FNAs; while very
specific, this test lacks sensitivity, as organisms may be
extremely sparse and difficult to visualize. Therefore, it has
become routine to submit needle rinses or separate needle
passes in appropriate media to microbiology, where direct
microscopy is performed, together with traditional culture.
If all tests are negative prior to culture becoming positive,
definitive diagnosis may be delayed for up to 6 weeks. In a
country or region in which MTB is so prevalent, treatment
delays of this nature are unacceptable, as they result in
continued presence of infected and infectious patients in
their communities, where further spread is likely.
To alleviate this, fluorescence, utilizing fixed Papanicolaou-stained cytologic material, becomes appealing for
rapid and cost-effective diagnosis of MTB. When judged
against traditional culture, the sensitivity in this study of
cytodiagnosis (84.9%), and that of autofluorescence of Papanicolaou-stained material (65.9%), may not reach levels
of sensitivity as attained in the diagnosis of malignancy by
FNA. These sensitivity values are, however, higher than
those quoted by Kupper et al.9 in their work on Mycobacterium kansasii. In terms of providing clinicians in overcrowded hospital settings with same-day working diagnoses
on which immediate antituberculous treatment can be commenced, it is our feeling that the routine addition of fluoDiagnostic Cytopathology, Vol 30, No 4
259
70
WRIGHT ET AL.
rescence to microscopy in FNAs has much to offer. It can be
regarded as a one-stop test, performed if necessary on one
glass slide, seen in one laboratory unit, with the final report
available within hours of the FNA procedure. Diagnosis is
accomplished without reprocessing from transport medium
for direct evaluation and for culture, as occurs in microbiology. It also avoids the use of radioactive substances that,
as indicated previously, carries some degree of morbidity
for laboratory workers.9 The percentage of cytology⫹/fluorescence⫹ patients subsequently found to be culture-negative (18.2% of those who might be placed on anti-MTB
therapy using this dual positivity as an indication; 8% of the
entire group referred for FNA) would appear to justify the
immediate commencement of such therapy, if clinically
consistent. Antituberculous therapy can be stopped if culture is
found to be negative, or if clinicoradiological assessment
deems the disease process not to be tuberculous in nature.
Culture was used in this study as the final arbiter of MTB
infection. However, while it is the most accurate means of
diagnosis available, it too is beset by an irreducible false
negative rate. It is thus possible that some patients who are
positive by cytology and/or fluorescence in this study, but
negative by subsequent culture, did in fact have tuberculosis. Only a long-term prospective clinicoradiologic and cytologic study incorporating review of patient records and
ultimate response to therapy over time could answer this
dilemma. An investigation of that nature is outside the
capabilities of this laboratory-based study.
There are ways in which the sensitivity of diagnostic
fluorescence could be improved. A recent study13 utilized
automated Papnet screening technology for the identification of small rare mycobacterial organisms stained by the
auramine-rhodamine method. However, for most laboratories, an innovation of that type would be financially unattainable. On the other hand, the mere addition of careful
dark-field illumination, searching for fluorescence as reported here, is feasible in most laboratories.
Compared with culture techniques, fluorescence yielded a
number of false positive cases in this study. Bacterial forms
other than MTB are known to exhibit autofluorescence; it is
thus possible that incorrect identification of these very small
organisms will be made. Kupper et al.8 has suggested that a
minimum number of eight organisms be required for positive diagnosis. In our experience in retroviral positive patients, MTB organisms are profuse and theoretically would
be easier to identify. The fluorescent staining properties of
Mycobacterium avium intracellulare are more variable that
than of MTB. In 596 culture-positive cases in this study,
only 6 were mycobacteria other than Mycobacterium tuberculosis. There is, therefore, no replacement for experience and
familiarity with the morphology of the organisms and scrupulous attention to detail when evaluating fluorescent bacteria.
It cannot be sufficiently emphasized that fluorescence of
organisms is appropriate only to material stained with the
260
Diagnostic Cytopathology, Vol 30, No 4
Papanicolaou stain. The results of the above study cannot,
for example, be transferred to cell block or histopathologic
sections stained with hematoxylin-eosin. The particular ingredient of the Papanicolaou stain that is thought to permit
autofluorescence is most likely EA50 or EA65. Because of
the capacity of the Papanicolaou stain to permit autofluorescence, cytopathologists find themselves in the unique
position of extending the usual capability of cytodiagnosis
to include identification of fluorescent organisms, such as
Mycobacterium tuberculosis, Pneumocystis, and Cryptococcus. In high-incidence areas of tuberculosis and of AIDS,
this method may prove to be diagnostically useful on a scale
larger than previously anticipated.
Acknowledgments
The authors thank Mrs. Lily Battaglia of the library at the
South African Institute for Medical Research (now National
Health Laboratory Service) for her ready and invaluable
assistance in locating scarce reports on the statistics of
communicable disease in South Africa.
References
1. Wilkinson D, Moore DA. HIV-related tuberculosis in South Africa:
clinical features and outcome. S Afr Med J 1996;86:64 – 67.
2. Connolly C, Davies GR, Wilkinson D. Impact of the human immunodeficiency virus epidemic on mortality among adults with tuberculosis in
rural South Africa, 1991–1995. Int J Tuberc Lung Dis 1998;2:919–925.
3. Del Amo J, Malin AS, Pozniak A, De Cock KM. Does tuberculosis
accelerate the progression of HIV disease? evidence from basic science and epidemiology. AIDS 1999;13:1151–1158.
4. Shinnink TM, Good RC. Diagnostic mycobacteriology laboratory
practices. Clin Inf Dis 1995;21:291–299.
5. Kommareddi S, Abramowski CR, Swinehart GL, Hrabak L. Nontuberculous mycobacterial infections: comparison of the fluorescent
auramine-0 and Ziehl-Neelsen techniques in tissue diagnosis. Hum
Pathol 1984;15:1085–1089.
6. Kumar N, Tiwari MC, Verma K. AFB staining in cytodiagnosis of
tuberculosis without classical features: a comparison of Ziehl-Neelsen
and fluorescent methods. Cytopathology 1998;9:208 –214.
7. Wright PW, Wallace RJ, Wright NW, Brown BA, Griffith DE. Sensitivity of fluorochrome microscopy for detection of Mycobacterium
tuberculosis versus nontuberculous mycobacteria. J Clin Microbiol
1998;36:1046 –1049.
8. Kupper TH, Steffen U, Wehle K, Richartz G, Pfitzer P. Morphological
study of bacteria of the respiratory system using fluorescence microscopy of Papanicolaou-stained smears with special regard to the identification of mycobacteria. Cytopathology 1995;6:338 – 402.
9. Kupper TH, Wehle K, Marzahn S, Pfitzer P. The cytologic diagnosis
of Mycobacterium kansasii tuberculosis by fluorescence microscopy of
Papanicolaou-stained specimens. Cytopathology 1995;6:331–338.
10. Zaharopoulos P. Demonstration of parasites in toxoplasma lymphadenitis by fine-needle aspiration cytology: report of two cases. Diagn
Cytopathol 2000;22:11–15.
11. Kocjan G, Miller R. The cytology of HIV-induced immunosuppression: changing pattern of disease in the era of highly active antiretroviral therapy. Cytopathology 2001;12:281–296.
12. Jannotta FS, Sidawy MK. The recognition of mycobacterial infections
by intraoperative cytology in patients with acquired immunodeficiency
syndrome. Arch Pathol Lab Med 1989;113:1120 –1123.
13. Veropoulis K, Learmonth G, Campbell C, Knight BK, Simpson J.
Automated identification of tubercle bacilli in sputum: a preliminary
investigation. Analyt Quant Cytol Histol 1999;21:277–282.
71
Diagnosing Mycobacterial
Lymphadenitis in Children Using
Fine Needle Aspiration Biopsy:
Cytomorphology, ZN Staining and
Autofluorescence—Making More of Less
Colleen A. Wright, F.R.C.Path.,1* Mirjam van der Burg, Ph.D.,1{ D. Geiger, M.Sc.,1
Jeroen G. Noordzij, Ph.D.,2{ Steven M. Burgess, Ph.D.,3 and Ben J. Marais, Ph.D.2
Although the incidence of TB has stabilized or declined in most
world regions, it is increasing in Africa, Southeast Asia, and the
Eastern Mediterranean, fuelled by the HIV pandemic. More than
4,000 people died daily from TB-related illnesses in 2005. TB is
a major cause of childhood morbidity and mortality in these
developing countries, and there is an urgent need for rapid and
definitive modalities for mycobacterial diagnosis in children.
This prospective study in Tygerberg Hospital, Cape Town, South
Africa, evaluates the ability of fine needle aspiration biopsy
(FNAB) to diagnose mycobacterial lymphadenitis in children,
using cytomorphology, autofluorescence on Papanicolaou
stained smears, Ziehl-Nielsen (ZN) staining and/or culture.
FNABs were performed on 200 children, and 25 (12.5%) aspirates were inadequate. Cultures were positive in 79/175 (45%);
Mycobacterium tuberculosis was identified in 61 and Mycobacterium bovis BCG in 18 cases. Using culture as the gold standard,
the concordance of the different techniques was as follows: cytomorphology 70%, ZN staining 73%, and autofluorescence 68%.
Using an alternative gold standard (culture positive and/or suggestive cytomorphology plus positive autofluorescence or ZN
smear), the ‘‘true’’ diagnostic performance of the various techniques was as follows: cytomorphology—sensitivity 78%, specificity 91%, positive predictive value (PPV) 93%, ZN staining - sensitivity 62%%, specificity 97%, PPV 97%; autofluorescence—
1
Division of Anatomical Pathology, Department of Pathology, University of Stellenbosch and NHLS Tygerberg Hospital, Tygerberg, South
Africa
2
Department of Pediatrics and Child Health and Ukwanda Centre for
Rural Health, University of Stellenbosch, South Africa
3
Graduate School of Business, University of Cape Town, Cape Town,
South Africa
{
Present address: Department of Immunology, Erasmus MC Rotterdam, The Netherlands.
{
Present address: Division of Pediatric Immunology and Infectious diseases, UMC St. Radboud, Nijmegen, The Netherlands.
*Correspondence to: Colleen A. Wright, F.R.C.Path., Discipline of
Anatomical Pathology, PO Box 19063, Tygerberg 7505, South Africa.
E-mail: cawr@sun.ac.za
Received 25 September 2007; Accepted 20 November 2007
DOI 10.1002/dc.20788
Published online in Wiley InterScience (www.interscience.wiley.com).
'
2008 WILEY-LISS, INC.
sensitivity 67%, specificity 97%, PPV 97%; and culture—sensitivity 75%, specificity 100%, and PPV 100%. FNAB was shown
to provide a rapid and definitive diagnosis in the majority of
cases of suspected tuberculous lymphadenitis in children, based
on cytomorphology and identification of the organism. Diagn.
Cytopathol. 2008;36:245–251. ' 2008 Wiley-Liss, Inc.
Key Words: fine needle aspiration biopsy; children; mycobacteria; tuberculosis; lymph node; HIV
Tuberculosis (TB) was declared a global emergency by
the World Health Organization (WHO) in 1993.1 Since
then, the incidence of TB has stabilized or declined in
most world regions, except Africa, Southeast Asia, and
the Eastern Mediterranean, where the impact of the pandemic is staggering.2 Between 2000 and 2005, new TB
cases increased worldwide by 6%, to more than 24,000
new TB cases daily. More than 84% of these cases were
located in sub-Saharan Africa and Asia. Even though TBrelated deaths declined by 11% between 2000 and 2005,
more than 4,000 people died daily from TB-related illnesses in 2005. It is fair to say that TB remains an emergency in those countries that can least afford the costs of
diagnosing and treating its victims.
TB is a major cause of childhood morbidity and mortality in developing countries.3,4 Children below 3 years of
age are particularly vulnerable because their immature
immune systems make them susceptible to more severe
disease manifestations, such as miliary TB and TB meningitis.5,6
In 1989, the WHO estimated the worldwide incidence
of TB in children to be 1.3 million per year with 450,000
children dying from TB-related illnesses in that year.7 In
2000 there were 8.3 million new cases of TB estimated
worldwide, with 11% occurring in children younger than
15 years of age.8 As alarming as these estimates are, they
Diagnostic Cytopathology, Vol 36, No 4
245
72
Diagnostic Cytopathology DOI 10.1002/dc
WRIGHT ET AL.
may grossly underestimate the incidence of TB in children because the WHO estimates are based on smear positive cases and more than 80% of children with TB are
not smear positive.9 In South Africa, as many as 40% of
all cases of TB occur in children.8
The diagnosis of TB in children is particularly challenging for three reasons.10–12 First, sputum and alternative specimens (e.g., gastric aspirates) are difficult to obtain and have a poor diagnostic yield (i.e., typically as
low as 30–40% even with more sensitive mycobacterial
culture).9 Second, culture has limitations. Although direct
culture inoculation at the bedside gives a high yield,13 it
takes 1–6 weeks to deliver a final result. Culture also fails
to differentiate between the various mycobacteria and,
thus, requires additional polymerase chain reaction
(PCR)-based testing for accurate speciation. Third,
although chest radiography is considered reasonably accurate and practical for diagnosis of intrathoracic TB in
children with suspicious symptoms, it has many limitations. Most importantly, it is not helpful in the diagnosis
of extrathoracic TB without concurrent intrathoracic manifestations, a combination common in children.2,10 For
example, a recent South African national study reported
that 39,739/270,178 (14.7%) of new TB cases in 2005
were extrapulmonary TB.2 A Cape Town study reported
that 72/439 (16.4%) of children treated for active TB had
extrathoracic disease and 65.4% of the positive cases had
no concurrent intrathoracic disease manifestations.11 TB
lymphadenitis was the most common form of extrathoracic TB in the Cape Town study and nearly all cases with
peripheral TB lymphadenitis presented with enlarged
cervical lymph nodes,11 consistent with previous
research.12,14 TB lymphadenitis is the most common
cause of persistent cervical lymphadenopathy in this TB
endemic area.13 The challenge of diagnosing TB in children highlights the urgent need for rapid and definitive
mycobacterial diagnosis in children with persistent cervical lymphadenopathy, which is the focus of our
research.
Fine-needle aspiration biopsy (FNAB) has many qualities that make it an excellent diagnostic procedure in
children with persistent superficial lymphadenopathy,
especially in TB-endemic countries with limited resources. FNAB is a simple technique to perform. It is minimally invasive, safe and has very limited side effects
when performed with a small gauge needle (e.g., no sinus
formation). It can be performed as an outpatient procedure by trained medical or paramedical staff.13,15 Collecting a representative sample of material from a lymph
node allows cytomorphological analysis and identification
of the organism by direct microscopy and/or culture.
The diagnostic value of various microscopy techniques
used to evaluate an FNAB sample has not been fully
evaluated in children. A pilot study of 65 aspirates in
246
Diagnostic Cytopathology, Vol 36, No 4
children using cytomorphology, ZN staining, and autofluorescence as diagnostic modalities had showed 92.3%
concordance with culture when all three modalities were
combined.16 The current study aimed to compare the
diagnostic performance of cytomorphology, ZN staining, autofluorescence, and culture in a large cohort of
children.
Materials and Methods
A prospective study was performed in Tygerberg Hospital
during the period of January 2003 to June 2005. All children (a) less than 13 years of age (eligible for access to
pediatric services), (b) with persistent superficial lymphadenopathy not responsive to first line antibiotic therapy,
and (c) with a clinical suspicion of mycobacterial infection routinely referred for diagnostic FNAB were included
in the study.
The referring clinician was requested to give oral sedation and analgesia 30 minutes prior to the procedure,
which was performed in the ward or outpatient clinic
with the assistance of the referring doctor and/or nursing
staff. Consent was obtained from the legal guardian
accompanying the child. The FNAB was performed using
a 23-g or 25-g needle attached to a 10-cc syringe. The
skin was cleaned using an alcohol swab, the child firmly
immobilized and the node stabilized by the aspirator
while performing the aspirate using no more than 2-cc
suction. Standard precautions were taken to minimize any
complications. Two smears were prepared from each aspirate and the needle and syringe rinsed in mycobacterial
growth tube indicator (MGIT, Beckton Dickinson, USA)
medium for mycobacterial culture. One smear was spray
fixed with commercial cytology fixative for Papanicolaou
staining and the other air-dried for Giemsa staining.
Smears were sent to cytology for cytomorphology, autofluorescence, and ZN staining.
The stained smears were screened and evaluated for adequacy and diagnosis using each of the modalities
described. Cytological results were reported as nonspecific
reactive lymphadenopathy or cytologically consistent with
mycobacterial infection, based on specific morphological
criteria which have been previously described.17–19 Immune competent patients with TB present with the more
classical morphological picture of epithelioid granulomata
and epithelioid histiocytes in a background of reactive
lymphocytes and plasma cells (Fig. C-1). A small amount
of amorphous necrosis and occasional giant cells may be
present. Patients with tuberculosis who are immune compromised have smears showing abundant ‘‘dirty’’ necrosis
in which neutrophils and cellular debris are prominent
(Fig. C-2). There is a spectrum of morphology between
these extremes, and aspirates from patients with mycobacterial lymphadenitis may occasionally show morphological features of suppurative acute lymphadenitis.
73
Diagnostic Cytopathology DOI 10.1002/dc
FINE NEEDLE ASPIRATION BIOPSY IN CHILDREN WITH MYCOBACTERIAL INFECTION
Fig. C-1–C-4. Fig C-1. Poorly formed granuloma consisting of epithelioid histiocytes and lymphocytes in a clean background, consistent with tuberculous lymphadenitis in an immune competent child (Papanicolaou, 3400). Fig. C-2. Numerous neutrophils and karyhorrectic debris in a necrotic background, consistent with tuberculous lymphadenitis in an immune compromised child. (Papanicolaou, 3400). Fig. C-3. Modified ZN stain in a lymph
node aspirate from an immune compromised child, demonstrating the overwhelming load of mycobacterial organisms present (Papanicolaou, 3400).
Fig. C-4. Autofluorescence of mycobacteria in a lymph node aspirate (Papanicolaou, 31,000 with a wide-band blue excitation filter).
In patients with lymphadenitis due to M bovis BCG,
histiocytes with abundant foamy cytoplasm are present.
Because of the considerable morphological variation, the
fact that patients with mycobacterial infection may present with aspirates showing nonspecific lymphadenitis and
that other organisms such as fungal infections may present a similar clinical and morphological picture, staining
for mycobacteria was performed an all aspirates, irrespective of the cytological diagnosis. ZN stains were performed on one of the Giemsa stained slides according to
a slightly modified technique. The smears were differentiated in 3% acid alcohol for 1 minute and counterstained
with 1% methylene blue for 10 seconds only (Fig. C-3).
One of the Papanicolaou stained slides was screened
using a Zeiss Axiophot microscope with a fluorescent
attachment and a wide-band blue excitation filter (450–
480 nm). Mycobacteria auto fluoresce as brilliant yellow
bacilli, thin and slightly curved with polar enhancement
and a uniform length of 2.0–2.7 lm (Fig. C-4).20 Despite
identification of the mycobacteria on ZN staining and/or
autofluorescence, Mycobacterium tuberculosis, Mycobacterium bovis BCG, and Mycobacterium avium–intracellulare are morphologically indistinguishable.
Culture samples were inoculated into 7H9 liquid growth
medium at the bedside and sent directly to microbiology,
according to standard protocol. Cultures were incubated at
378C for 8 weeks or until indicated as positive, according
to standard protocol using the MGIT system (Becton Dickenson, Cockeysville, MD). All mycobacterial isolates were
identified as Mycobacterium tuberculosis, Mycobacterium
bovis BCG or nontuberculous mycobacteria (NTM) by polymerase chain reaction (PCR) testing.
Diagnostic Cytopathology, Vol 36, No 4
247
74
Diagnostic Cytopathology DOI 10.1002/dc
WRIGHT ET AL.
Parents gave written informed consent for study participation and ethics approval was obtained from the Institutional Review Board of Stellenbosch University.
is not uncommon. The previous pilot study investigated
the culture negative cases in which any of the other
modalities were positive and identified eight false culture
negative cases that were clinically assessed as having TB
and responded to TB therapy.16 Patients may commence
TB treatment prior to referral for FNAB. Therefore, using
culture alone as the gold standard may be flawed, and we
identified an alternative gold standard as culture positive
or positive cytomorphology plus positive autofluorescence
or ZN staining.
Using this alternative gold standard improved the diagnostic performance of all three modalities. The concordance rate with the new standard is 83% for cytomorphology, 79% for autofluorescence, 76% for ZN staining, and
84% for culture (Table III). The sensitivity, specificity,
and overall efficacy of these diagnostic modalities are
shown in Table IV.
Results
FNABs were collected from 200 children, and 25 (12.5%)
aspirates were inadequate. Samples from 175 children
were included in the analysis. Patient demographics and
sample data are reported in Table I.
The results for the three diagnostic techniques were as
follows. Cultures were positive in 79 cases (45%): 61
Mycobacterium tuberculosis, 18 Mycobacterium bovis
BCG and no NTM positive samples. The median age of
children with positive cultures was similar to those with
negative cultures (2 years). However, all children with
Mycobacterium bovis BCG were less than 2 years of age,
the majority 16/18 (89%) being infants.
Table II compares the diagnostic performance of the
three microscopy techniques cytomorphology, autofluorescence, and ZN staining alone and in combination to culture as the gold standard. However, culture is probably
not the optimal gold standard, as a false negative culture
Discussion
In many low- and middle-income countries with limited
resources, the diagnosis of TB is still based on poorly
validated symptom-based algorithms, often not resulting
in a definitive diagnosis.21 FNAB provides a rapid and definitive tissue diagnosis in the majority of children with
superficial lymphadenopathy. This study demonstrates that
it also permits confirmation of the presence of mycobacteria, with fluorescence microscopy being more sensitive
than ZN staining. Recent studies have demonstrated that
light emitting diode (LED) technology provides a cheap
and reliable light source with minimal energy requirements that performs as well as the traditional mercury
vapour lamp used in this study.22,23 To date, the high cost
and short half-life of the mercury vapour lamp has limited
the availability of fluorescence microscopy in healthcare
systems with constrained financial resources.23
The 2007 WHO report noted that over half of the populations in the African, South-East Asia, and Western
Pacific regions had limited coverage of culture services.
Most countries had neither national policies to expand
culture and sensitivity testing services nor the technical
capacity to implement and support such services. In South
Africa., there were 143 laboratories performing smear microscopy and 18 laboratories performing culture.2 This is
mainly due to centralization of the more expensive and
Table I. Demographics and Sample Characteristics
Numbers
Percentage
200
25
175
12.5
87.5
75
48
32
20
42.9
27.4
18.3
11.4
82
93
46.9
53.1
0
86
89
73
68
79
–
49.1
50.1
41.7
38.9
45.0
61
18
0
77.0
23.0
–
Number of aspirates
Inadequate samples
Specimens included in analysis
Age
1 year
2–4 years
5–9 years
10 years
Gender
Male
Female
Diagnosis
Malignancy
Nonspecific node l
Consistent with mycobacterial infection
Fluorescence positive
ZN positive
Culture positive mycobacterial infection
Organism isolated
M. Tuberculosis
M. bovis BCG
Nontuberculous mycobacteria
Table II. Concordance of Diagnostic Modalities with Culture as Gold Standard
Cytology
Positive agreement
Negative agreement
Concordant
False positive
False negative
Pearson v2
Phi
Fluorescence
ZN
Any diagnostic modality
Any two diagnostic modalities
27%
41%
68%
14%
18%
21.48
0.35
29%
45%
73%
10%
17%
36.19
0.46
39%
34%
73%
21%
6%
43.21
0.50
30%
39%
69%
15%
15%
24.86
0.38
33%
37%
70%
18%
12%
29.33
0.41
n ¼ 175; all Pearson v2 statistics are significant at P 0.0001.
248
Diagnostic Cytopathology, Vol 36, No 4
75
Diagnostic Cytopathology DOI 10.1002/dc
FINE NEEDLE ASPIRATION BIOPSY IN CHILDREN WITH MYCOBACTERIAL INFECTION
Table III. Concordance of Diagnostic Modalities with New Gold
Standard Culture Positive or Cytomorphology Positive Plus Positive
Autofluorescence or ZN
Positive agreement
Negative agreement
Concordant
False positive
False negative
Pearson v2
Phi
Cytology
Fluorescence
ZN
Culture
47%
36%
83%
3%
13%
81.02
0.68
41%
38%
79%
1%
20%
70.59
0.64
38%
38%
76%
1%
23%
61.99
0.60
45%
39%
84%
0%
15%
93.74
0.73
n ¼ 175; all Pearson v2 statistics are significant at P 0.0001.
sophisticated culture techniques. Centralization is also a
feasible option in resource constrained healthcare systems
because FNAB can be done as an outpatient procedure by
trained medical or paramedical personnel and correctly
fixed slides can be sent to specialized diagnostic facilities
for cytological evaluation and fluorescent microscopy.
The diagnosis of TB lymphadenitis is increasingly complicated by the HIV pandemic. HIV infection, in children
and adults, is a major contributor to the escalating TB
epidemic worldwide.24–26 At the end of 2000, it was estimated that there were 11.5 million HIV-infected people
coinfected with TB in sub-Saharan Africa.1 In 2005,
South Africa with 0.7% of the world’s population had
19% of all cases of TB in adult HIV-positive people.2
Infection with the HIV virus is estimated to increase the
lifetime risk of a patient infected with M tuberculosis of
developing TB by 10-fold to 50%.1 Children are not
exempt from this burden.
A hospital-based study from Zambia demonstrated a
HIV seroprevalence rate in children with TB to be >70%,
while the minority of childhood TB cases were HIV negative.27 In HIV-infected children the diagnosis of intraand/or extrathoracic TB is even more difficult due to concomitant chronic diseases such as lymphocytic interstitial
pneumonitis (LIP) and other opportunistic infections.27,28
Persistent peripheral lymphadenopathy in children is common and is itself a criterion for the classification of HIVrelated disease in children.29,30 In HIV-infected children,
peripheral lymphadenopathy may be associated with
many pathological entities, from infections to neoplasia.
A study from South Africa demonstrated that the majority of HIV-infected children with persistent lung disease (PLD), due to a variety of causes, have persistent
peripheral lymphadenopathy (PPL) and PPL was present
in many children with pulmonary TB (PTB).31 Thus,
enlarged peripheral lymph nodes were identified as important specimen collection sites to consider in establishing a
definitive diagnosis of TB in HIV-infected children.31 In
HIV-infected children, it is even more important to establish a mycobacteriologic diagnosis, as tuberculin skin
tests, symptom-based approaches, and radiology are less
Table IV. Summary of All Diagnostic Modalities Using an Alternative
Gold Standard: Culture Positive or Cytomorphology Positive Plus
Positive Autofluorescence or ZN
Sensitivity
Specificity
PPV
NPV
Efficacy
Cytology (%)
Fluorescence (%)
ZN (%)
Culture (%)
78
91
93
73
83
67
97
97
66
79
62
97
97
63
76
75
100
100
72
84
helpful,28 and these children may be at an increased risk
of exposure to drug resistant strains of TB.
Disseminated BCG disease is a rare complication of
BCG vaccination, classically only described in children
with severe congenital immune deficiencies. However,
HIV-infected children also appear to be at high risk.32,33
BCG vaccination may also associated with local complications such as abscess formation at the vaccination site
and/or regional lymphadenitis, with reported rates of
lymphadenitis in BCG recipients ranging from 0.5% to
17.6%.24 In South Africa, the recent change in vaccine
policy in July 2000 from percutaneous Tokyo strain BCG
to intradermal Danish strain BCG vaccine has seen a
number of adverse events reported in both HIV-infected
and immune competent children.32 In children with axillary or other regional lymph node enlargement, FNAB
offers a simple first line diagnostic procedure to confirm a
diagnosis of BCG disease.
FNAB is widely utilized in the diagnosis of palpable
masses including peripheral lymphadenopathy, and its
value in the diagnosis of mycobacterial lymphadenitis in
adults is well documented.17–20,34,35 It offers a simple,
effective, and safe modality for obtaining a representative
sample of material from a lymph node, permitting cytological evaluation, identification of the organism by morphology, culture and molecular techniques such as the
polymerase chain reaction (PCR). Cytomorphology is
simple, but not specific, as other opportunistic infections
may present with similar cytological changes and the
quality of the smear is operator dependent.17,19 With good
clinical triage of patients, its sensitivity and specificity
can be increased. In our study, cytomorphology achieved
a sensitivity of 78% and a specificity of 91%.
ZN staining, even in good laboratories, is generally
regarded as insensitive, identifying organisms in only
about 20% of culture positive samples.35 This may be
dependent on experience, the staining, and the bacterial
load. In our study, ZN staining achieved sensitivity of
62% but with a specificity of 97%. Although autofluorescence is simple, sensitive, and inexpensive, it is not
widely used. It requires a fluorescent microscope, which
may not be readily available, but has the advantage of not
requiring additional stains and is therefore inexpensive
and rapid. In our study, it performed similar to ZN stainDiagnostic Cytopathology, Vol 36, No 4
249
76
Diagnostic Cytopathology DOI 10.1002/dc
WRIGHT ET AL.
ing in identifying mycobacterial infection with a sensitivity of 67% and a specificity of 97%. Its specificity is
dependant on the experience of the operator and, as ZN
staining, it cannot differentiate between the various mycobacteria.20 These results may be partly due to the high
incidence of mycobacterial disease in South Africa and
good clinical selection of patients.
Culture, where feasible and affordable, is an important
part of the investigation of suspected mycobacterial
lymphadenitis. Direct inoculation into MGIT or BactecTM
medium at the bedside gives a high yield,14 but it takes
1–6 weeks for a result, and fails to differentiate the various mycobacteria unless additional PCR-based testing is
available for accurate speciation of M. tuberculosis complex. Culture has inherent problems and may be influenced by many factors such as inefficient sampling, contamination, variable bacterial load, and commencement of
therapy prior to the diagnostic procedure. If the cytomorphology is consistent with mycobacterial infection, and
the organism is identified by ZN staining or autofluorescence, the probability of a false positive diagnosis is
small, and patients may safely commence therapy. FNAB
of superficial lymph nodes is an outpatient procedure and
requires little infrastructure and equipment and is therefore ideal for resource-limited countries, which are the
very same countries that bear the burden of HIV and TB.15
The basic diagnostic modalities of cytomorphology and
subsequent morphological identification of the organism
are readily available and relatively inexpensive. Although
culture adds to the cost of the diagnosis, its value lies in
providing an opportunity for speciation and sensitivity
testing, as well as surveillance of prevalent strains within
a community. Although radiometric methods such a
MGIT have reduced the time to culture mycobacterial
organisms, this may still be unacceptably long for patients
who have difficulty in returning for results. Other innovative methods for diagnosing TB are being developed but
remain mainly experimental and are still relatively expensive.21
Most children referred with suspected mycobacterial
lymphadenitis have a tuberculin skin test performed and
must return to the clinic for this to be read with in 48–72
hours. If FNAB is performed at the time of the first visit,
the results of the aspirate can be available at the return
visit 48 hours later. This enables children to be put on
appropriate therapy with a high degree of certainty, pending the results of culture and speciation. In countries
where culture is not available, FNAB provides the clinician with the ability to make a definitive diagnosis based
on cytomorphology and identification of the organism,
particularly in immune compromised children, who
require urgent and appropriate treatment. If cytology services are not available, children with suspected mycobacterial lymphadenopathy may have an FNAB performed and
250
Diagnostic Cytopathology, Vol 36, No 4
smears made for direct TB microscopy. Although this is
not ideal, it provides reasonable diagnostic accuracy, with
the additional option to courier-fixed slides to a center
with adequate cytology services. This will help health
care workers in TB-endemic countries with limited
resources to navigate the diagnostic minefield of persistent
lymphadenopathy in children.
In conclusion, FNAB proved highly feasible and provided a rapid and definitive diagnosis in the majority of
cases, irrespective of the diagnostic modality used.
References
1. Harries A, Maher D, Graham S. TB/HIV: A clinical manual. Geneva: World Health Organization; 2004. WHO/HTM/TB/2004.329.
2. World Health Organization. Global tuberculosis control: Surveillance, planning, financing. Geneva: World Health Organization;
2007. WHO Report 2007, WHO/HTM/TB/2007.376.
3. Beyers N, Gie RP, Schaaf HS, Van Zyl S. A prospective evaluation
of children under the age of five years living in the same household
as adults with recently diagnosed pulmonary tuberculosis. Int J
Tuberc Lung Dis 1997;1:38–43.
4. Starke JR. Tuberculosis in children. In: Reichman LB, Herschfield
ES, editors. Tuberculosis. A comprehensive international approach.
New York: Marcel Dekker; 1993. p 329–367.
5. Marais BJ, Gie RP, Schaaf HS, et al. The natural history of childhood intra-thoracic tuberculosis: A critical review of literature
from the pre-chemotherapy era. Int J Tuberc Lung Dis 2004;8:392–
402.
6. Marais BJ, Donald PR, Gie RP, Schaaf HS, Beyers N. Diversity of
disease manifestations in childhood pulmonary tuberculosis. Ann
Trop Paediatr 2005;25:79–86.
7. Kochi A. The global tuberculosis situation and new control strategy
of the WHO. Tubercle 1991;72:1–6.
8. Nelson LJ, Wells CD. Tuberculosis in children: Considerations for
children from developing countries. Semin Pediatr Infect Dis
2004;15:150–154.
9. Walls T, Shingadia D. Global epidemiology of paediatric tuberculosis. J Infect 2004;48:13–22.
10. Theart AC, Marais BJ, Gie RP, Hesseling AC, Beyers N. Criteria
used for the diagnosis of childhood tuberculosis at primary health
care level in a high-burden, urban setting. Int J Tuberc Lung Dis
2005;9:1210–1214.
11. Marais BJ, Gie RP, Schaaf HS, Hesseling AC, Enarson DA, Beyers
N. The spectrum of disease in children treated for tuberculosis in a
highly endemic area. Int J Tuberc Lung Dis 2006;10:732–738.
12. Miller FGW, Seale RME, Taylor MD. Tuberculosis in children.
Boston: Little Brown; 1963.
13. Marais BJ, Wright CA, Schaaf HS, et al. Tuberculous lymphadenitis as a cause of persistent cervical lymphadenopathy in children
from a tuberculosis-endemic area. Pediatr Infect Dis J 2006;25:142–
146.
14. Kumar RK. Tuberculous lymphadenitis in children—Role of fine
needle aspiration cytology. J Assoc Physicians India 1999;47:976–
999.
15. Wright CA, Pienaar JPP, Marais BJ. Fine needle aspiration biopsy
in children—Diagnostic utility in resource-limited settings. Ann
Trop Paedtr 2008;28 (in press).
16. Wright CA, Burgess SM, Geiger D, Wasserman E, van den Burg
M. The diagnosis of mycobacterial lymphadenitis in children: Is
fine needle aspiration the way to go? In: The 46th Conference of
the Federation of South African Societies of Pathology, Durban,
South Africa, 2006.
77
Diagnostic Cytopathology DOI 10.1002/dc
FINE NEEDLE ASPIRATION BIOPSY IN CHILDREN WITH MYCOBACTERIAL INFECTION
17. Kocjan G, Miller R. The cytology of HIV-induced immunosuppression. Changing pattern of disease in the era of highly active antiretroviral therapy. Cytopathology 2001;12:281–296.
18. Purohit SD, Purohit V, Mathur ML. A clinical scoring system as
useful as FNAC in the diagnosis of tuberculous lymphadenitis in
HIV positive patients. Curr HIV Res 2006;4:459–462.
19. Nayak S, Mani R, Kavatkar AN, Puranik SC, Holla VV. Fine-needle aspiration cytology in lymphadenopathy of HIV-positive patients.
Diagn Cytopathol 2003;29:146–148.
20. Wright CA, van Zyl Y, Burgess SM, Blumberg L, Leiman G.
Autofluorescence of mycobacteria on lymph node aspirates—A
glimmer in the dark? Diagn Cytopathol 2004;30:257–260.
21. Marais BJ, Pai M. Recent advances in the diagnosis of childhood
tuberculosis. Arch Dis Child 2007;92:446–452.
22. Anthony RM, Kolk AH, Kuijper S, Klatser PR. Light emitting
diodes for auramine O fluorescence microscopic screening of Mycobacterium tuberculosis. Int J Tuberc Lung Dis 2006;10:1060–1062.
23. Hung NV, Sy DH, Anthony RM, Cobelens FGJ, van Soolingen D.
Fluorescence microscopy for tuberculosis diagnosis. Lancet Infect
Dis 2007;7:238–239.
24. Braun MM, Cauthen G. Relationship of the human immunodeficiency virus epidemic to pediatric tuberculosis and Bacillus Calmette-Guerin-vaccination. Pediatr Infect Dis J 1992;11:220–227.
25. Jeena PM, Mitha T, Bamber S, Wesley A, Coutsoudis A, Coovadia
HM. Effects of human immunodeficiency virus on tuberculosis in
children. Tuberc Lung Dis 1996;77:437–443.
26. Madhi SA, Huebner RE, Doedens L, Aduc T, Wesley D, Cooper
PA. HIV-1 coinfection in children hospitalized with tuberculosis in
South Africa. Int J Tuberc Lung Dis 2000;4:448–454.
27. Graham SM, Coulter JB, Gilks CF. Pulmonary disease in
HIV—Infected African children. Int J Tuberc Lung Dis 2001;5:
12–23.
28. Marais BJ, Cotton M, Graham S, Beyers N. Diagnosis and management challenges of childhood TB in the era of HIV. J Infect Dis
2007;196 (Suppl 1):S76–S85.
29. World Health Organization. Provisional guidelines for the diagnosis and classification of the EPI target diseases for primary health
care, surveillance and special studies. Geneva: WHO; 1983. EPI/
GEN/83/4.
30. Centers for Disease Control and Prevention. 1994. Revised classification system for human immunodeficiency virus infection in children less than 13 years of age. MMWR 1994;43;RR-12.
31. Jeena PM, Coovadia HM, Hadley LG, Wiersma R, Grant H, Chrystal V. Lymph node biopsies in HIV infected and noninfected children with persistent lung disease. Int J Tuberc Lung Dis 2000;4:
139–146.
32. Hesseling AC, Rabie H, Marais BJ, et al. Bacille Calmette-Guerin
vaccine-induced disease in HIV-infected and HIV-uninfected children. Clin Infect Dis 2006;42:548–558.
33. Hesseling AC, Marais BJ, Gie RP, et al. The risk of disseminated
Bacille Calmette-Guerin (BCG) disease in HIV-infected children.
Vaccine 2006;25:14–18.
34. Ersoz C, Polat A, Serin MS, Soylu L, Demircan O. Fine needle
aspiration (FNAB) cytology in tuberculous lymphadenitis. Cytopathology 1998;9:201–207.
35. Gupta AK, Nayar M, Chandra M. Critical appraisal of fine needle
aspiration cytology in tuberculous lymphadenopathy. Acta Cytol 1992;
36:391–364.
Diagnostic Cytopathology, Vol 36, No 4
251
78
CHAPTER 6
OPTIMISING CULTURE-BASED AND MOLECULAR
DIAGNOSIS
Mycobacterial Transport medium for routine culture of Fine Needle Aspiration Biopsies
CA Wright, C Bamford, Y Prince, A Vermaak, KGP Hoek, BJ Marais, RM Warren,
Archives of Disease in Childhood: in press
Combining Fine Needle Aspiration Biopsy (FNAB) and High Resolution Melt Analysis
to reduce diagnostic delay in Mycobacterial Lymphadenitis
CA Wright, KGP Hoek, BJ Marais, P van Helden, RM Warren,
Diagnostic Cytopathology: in press
79
Mycobacterial Transport medium for routine culture of Fine Needle Aspiration Biopsies
CA Wright,* C Bamford,† Y Prince,† A Vermaak,* KGP Hoek,‡ BJ Marais,§ RM Warren,‡
*
Divisions of Anatomical Pathology and †Medical Microbiology, Department of Pathology,
Stellenbosch University and NHLS Tygerberg Hospital, Tygerberg, South Africa
‡
NRF Centre of Excellence in Biomedical Tuberculosis Research / MRC Centre for
Molecular and Cellular Biology, Division of Molecular Biology and Human Genetics,
Stellenbosch University, South Africa
§
Department of Pediatrics and Child Health and Desmond Tutu TB Centre, Stellenbosch
University, South Africa
SUMMARY
Fine needle aspiration biopsy (FNAB) offers a simple outpatient technique for specimen
collection in paediatric tuberculosis (TB) suspects with peripheral lymphadenopathy. Culture
facilities are usually centralized; to perform FNAB with mycobacterial culture on an
outpatient basis requires use of a sterile transport medium to facilitate bedside inoculation,
maintain organism viability and reduce contamination risk en route to the laboratory.
We compared the mycobacterial yield and time to positive culture following bedside
inoculation into standard MGIT tubes versus initial inoculation into an inexpensive “in house”
liquid growth medium used for transport to the laboratory followed by immediate and delayed
MGIT inoculation (laboratory inoculation).
Over the period of one year 150 FNAB’s were performed; 57 (38%) cultured M. tuberculosis
complex. There was one case each of NTM and M bovis BCG, with the remaining 55 being M
tuberculosis. Results were concordant in 142 (94.7%) bedside and laboratory inoculation
pairs. There was no significant difference in time to positive culture between the bedside and
laboratory inoculation tubes (16.2 days S.D. 0.87 versus 17.1 days S.D. 0.85). There were 31
80
pairs with matched immediate and delayed laboratory inoculation culture results, of which 29
were concordant (93.6 %). The 2 discordant pairs showed no growth in the delayed culture;
both had required repeated decontamination due to bacterial contamination introduced in the
laboratory.
The use of inexpensive “in house” liquid growth medium transport bottles, combined with
practical tuition in FNAB, will improve cost effective diagnosis of TB at primary health care
level.
INTRODUCTION
Tuberculosis (TB) may be extrapulmonary in 10-30% children,1 and TB lymphadenitis is the
most common extra-pulmonary manifestation of TB in endemic areas;2 5-10% of children
may have TB adenitis in association with pulmonary involvement. 3 This provides an
excellent opportunity to obtain bacteriologic confirmation using fine needle aspiration biopsy
(FNAB).4 FNAB is a simple and safe outpatient procedure that can be performed by nurses in
resource-limited settings5 and provides material for direct microscopy as well as culture and
susceptibility testing. Cultures from FNAB specimens in children give a greater yield in a
shorter time, compared to conventional respiratory specimens such as gastric washings.6
Direct bedside inoculation of FNAB specimens is ideal and excellent results have been
obtained with commercial liquid media systems such as Mycobacterial Growth Indicator
Tubes (MGIT, Beckton Dickinson, USA).4 This practice however, is limited by cost and
availability.4 We aimed to develop a simple and cheap mycobacterial transport medium and
then to compare the bacteriologic yield and time to positive culture achieved with direct
bedside inoculation into MGIT versus initial use of the transport medium.
81
MATERIALS AND METHODS
All adults and children referred to the FNAB Clinic at Tygerberg Hospital, Cape Town, South
Africa (June 2007-May 2008) with clinical mycobacterial lymphadenitis and in whom written
informed consent to participate in the study was obtained (following ethics approval) were
included.
TB transport bottles, containing 1 ml aliquots of Middlebrook 7H9 broth (with 0.2% glycerol
and 0.05% Tween 80 added) in 10ml headspace glass vials, were prepared “in-house”
(Biomedical TB Research Unit by a study scientist, KGPH) in a laminar flow cabinet, sealed
with 20mmTFE/Sil Septa and 20mm Aluminium open top seals and autoclaved at 120º C /
20min (Figure 1). Unlike the MGIT tube, the transport bottle does not require removal of the
lid, minimizing the risk for contamination.
FNAB was done following standard protocol.4 Two needle passes were performed: from the
first, cytology slides were prepared before the syringe and needle were rinsed by withdrawing
an aliquot of liquid growth media into the syringe and then expelling the contents back into
the MGIT tube (bedside inoculation or control). After transport to the laboratory within the
same hospital, within 2-24 hours, the MGIT tube was incubated in a BACTEC MGIT 960
machine for 42 days.
From the second FNAB pass, slides were prepared for cytology and the syringe and needle
were rinsed into the TB transport bottle. In the laboratory, 0.5ml was aspirated from the
transport bottle, inoculated into a separate MGIT tube (immediate laboratory inoculation) and
incubated as above. In a subset of consecutive samples, the transport bottle was held at room
temperature (12-30 deg Celsius) for 7 days after which subculture was performed as described
82
above (delayed laboratory inoculation). The delayed inoculation group was designed to mimic
the situation of rural clinics where specimens may take some days to reach the laboratory.
PANTA antibiotic mixture (containing polymixin B, amphotericin B, nalidixic acid,
trimethoprim and azlocillin) (Becton Dickinson, New York, NY, USA) was added to all
MGIT tubes before incubation. All positive MGIT tubes were confirmed to contain acid fast
bacilli in the absence of bacterial contamination, followed by accurate polymerase chain
reaction (PCR) speciation .7 The mycobacterial yield and time to positive culture in the
respective MGIT tubes were compared using Statistica Version 8 software and the Chi2 test.
RESULTS
Over the period of one year 150 FNAB’s were performed, and 24.5% of these were in patients
less than 16 years, the youngest being 5 months old. M. tuberculosis complex was cultured in
57 patients (38%). There was one case each of NTM and M bovis BCG, with the remaining 55
being M tuberculosis; consistent with the profile of disease in our population.3, 4, 8 In the TB
culture negative patients there were an additional 12 cases (11.4%) positive on cytology
(morphological evidence of mycobacterial infection and identification of the organism). In 17
patients there was a specific diagnosis other than mycobacterial infection, made on cytology,
13 of these malignant neoplasms.
Results of culture were concordant in 142 (94.7%) bedside and laboratory inoculation pairs,
with 8 discordant pairs: 3 were positive on bedside inoculation only and 5 were positive on
laboratory inoculation only. In one of the discordant pairs, positive on bedside inoculation
and negative on laboratory inoculation there was inadequate material on the cytology slides.
There was no significant difference in time to positive culture between the bedside and
laboratory inoculation tubes (16.2 days S.D. 0.87 versus 17.1 days S.D. 0.85).
83
There were 31 pairs with matched immediate and delayed laboratory inoculation culture
results, of which 29 were concordant (93.6 %). The 2 discordant pairs showed no growth in
the delayed culture, compared to growth in bedside and immediate laboratory cultures; both
had required repeated decontamination due to bacterial contamination introduced in the
laboratory. There was no significant difference in time to positive culture between the
immediate and delayed laboratory inoculation groups (Table 1).
DISCUSSION
Although FNAB has been shown to be an excellent method for obtaining specimens for
culture as well as cytology, the method for transporting specimens to the laboratory is rarely
specified.9 Mycobacteria need to remain viable during transportation if culture is to be
successful and traditional methods for the preservation of specimens , e.g. refrigeration10 are
not applicable to FNAB specimens in resource- poor settings. In most patients FNAB yields
insufficient material to express into a sterile tube, and commercial liquid culture media, e.g.
BACTECTM MYCO/F LYTIC are recommended only for blood and body fluids.11
Submission of the syringe (with or without the needle attached) directly to the laboratory
compromises the quality of the specimen and puts health care workers at risk. This practice is
unacceptable, especially in settings with a high prevalence of MDR TB and/or HIV.
Rinsing the needle and syringe in commercial liquid culture medium, such as MGIT, is very
successful in obtaining a positive culture.4, 5 However, the widespread distribution of MGIT
tubes is limited by issues of cost and storage requirements. These tubes need to be stored in
the dark and removal of the cap for inoculation outside the laboratory increases the risk for
contamination. This study suggests that subculture from a TB transport bottle, even after a 7
day delay, gives similar yields in a similar time period, to direct bedside inoculation. The
84
few discrepancies (8/150) noted between the bedside and laboratory cultures may be due to
random sampling differences as the tubes were inoculated from separate needle passes.
The transport medium described in this study is inexpensive (MGIT tube $4 vs. $1 TB
transport bottle, 2008 prices in South Africa), easy to prepare “in house”, does not require
refrigeration as it is stable through a range of temperatures, and does not require removal of
the lid for inoculation. Transport of these specimens from rural clinics may follow the same
route as other laboratory specimens such as sputum and peripheral blood. The positive
culture yield and the time to positive culture from the transport medium, even after 7 days at
room temperature, are statistically comparable to direct bedside inoculation into MGIT tubes.
The production and distribution of these bottles to clinics and hospitals, combined with
practical tuition in the optimal performance of FNAB, will have a positive impact on the cost
effective diagnosis of TB at primary health care level.
ACKNOWLEDGEMENTS
Professor Martin Kidd (Stellenbosch University) for statistical analysis. This study is in
partial fulfillment of a PhD thesis.
CA Wright was funded by grant from the National Health Laboratory Services, South Africa.
Competing interests: none declared
Copyright licence statement
The Corresponding Author has the right to grant on behalf of all authors and does grant on
behalf of all authors, an exclusive licence (or non-exclusive for government employees) on a
worldwide basis to the BMJ Publishing Group Ltd and its Licensees to permit this article (if
accepted) to be published in Archives of Disease in Childhood and any other BMJPGL
85
products to exploit all subsidiary rights, as set out in our licence
(http://adc.bmj.com/iforalicence.pdf).
What is already known about this topic.
FNAB is used as to diagnose suspected mycobacterial lymphadenopathy.
Mycobacteria may be seen on staining FNAB material, but this does not allow speciation (TB
or non-tuberculous mycobacteria) or antimicrobial sensitivity testing.
What this study adds
A simple and inexpensive means to obtain mycobacterial culture from FNAB specimens was
developed using an “in house” liquid growth medium for transport to the laboratory.
Transporting FNAB material to the laboratory in the “in house” medium was similar to a
commercial medium with respect to culturing mycobacteria and in time to positive culture.
Similar results were obtained when the ‘in house” transport bottle was kept at room
temperature for 7 days before culturing for mycobacteria.
REFERENCES
[1]
Cruz AT, Starke JR. Clinical manifestations of tuberculosis in children. Paediatr
Respir Rev. 2007;8(2):107-17.
[2]
Marais BJ, Gie RP, Schaaf HS, et al. The spectrum of childhood tuberculosis in a
highly endemic area. Int J Tuberc Lung Dis. 2006;10:732-8.
[3]
Marais BJ, Wright CA, Schaaf HS, et al. Tuberculous Lymphadenitis as a Cause of
Persistent Cervical Lymphadenopathy in Children From a Tuberculosis-Endemic Area.
Pediatr Infect Dis J. 2006;25:142-6.
[4]
Wright CA, van der Burg M, Geiger D, et al. Diagnosing Mycobacterial lymphadenitis
in children using Fine Needle Aspiration Biopsy: Cytomorphology, ZN staining and
Autofluorescence – making more of less. Diagn Cytopathol. 2008;36:245-51.
86
[5]
Wright CA, Pienaar JP, Marais BJ. Fine needle aspiration biopsy: diagnostic utility in
resource-limited settings. Ann Trop Paediatr. 2008; 28:65-70.
[6]
Wright C, Hesseling A, Marais B. A comparison of Fine Needle Aspiration Biopsy
and other specimen collection methods in the diagnosis of Paediatric Tuberculosis. 48th
Conference of the Federation of South African Societies of Pathology. Cape Town, South
Africa 2008.
[7]
De Wit D, Steyn L, Shoemaker S, Sogin M. Direct detection of Mycobacterium
tuberculosis in clinical specimens by DNA amplification. J Clin Microbiol. 1990;28:2437-41.
[8]
Wright CA, van Zyl Y, Burgess SM, Blumberg L, Leiman G. Auto fluorescence of
Mycobacteria on Lymph Node Aspirates – A glimmer in the dark? Diagn Cytopathol.
2004;30:257-60.
[9]
Hall GS. Primary processing of specimens and isolation and cultivation of
mycobacteria. Clin Lab Med. 1996;16:551-66.
[10]
Kent P, Kubica G. Public Health Mycobacteriology A guide for the Level III
laboratory. Atlanta: USDHHS, Centers for Disease Control 1985.
[11]
Becton Dickinson Microbiology Systems. BACTEC™ Myco/F Lytic Medium 2009
[cited 02/08/09]; Available from: http://www.bd.com/ds/productCenter/442288.asp
87
Table 1
The mycobacterial yield and time to positive culture achieved with fine needle aspiration
biopsy and variable timing of MGIT* inoculation
Timing of MGIT*
inoculation
Time to positive culture
Mycobacterial culture
Total no
specimens
Positive
Negative
Mean
days
Standard
deviation
95% CI#
Bedside vs. immediate laboratory inoculation from transport bottle
Bedside
150
53 (35.3%)
97 (64.7%)
16.2
0.87
16.8 - 20.3
Immediate laboratory
150
55 (36.7%)
95 (63.3%)
17.1
0.85
15.7-19.1
Subset - matched pairs: delayed vs. immediate laboratory inoculation from transport bottle
Delayed laboratory
(day 7)
Immediate laboratory
(day 0)
31
11 (35.5%)
20 (64.5%)
17.1
1.36
14.1-20.1
31
13 (41.9%)
18 (58.1%)
16.2
1.31
13.3-19.2
*MGIT- Mycobacterial Growth Indicator Tube
#
CI- Confidence Interval
88
Figure 1. TB Transport Bottle
89
Combining Fine Needle Aspiration Biopsy (FNAB) and High Resolution Melt Analysis
to reduce diagnostic delay in Mycobacterial Lymphadenitis
CA Wright,1* KGP Hoek,2 BJ Marais,3 P van Helden,2 RM Warren,2
Division of Anatomical Pathology ,Department of Pathology, Stellenbosch University and
NHLS Tygerberg, South Africa; NRF Centre of Excellence in Biomedical Tuberculosis
Research / MRC Centre for Molecular and Cellular Biology, Division of Molecular Biology
and Human Genetics, Stellenbosch University, South Africa; 2 and Department of Pediatrics
and Child Health, Stellenbosch University, South Africa.
ABSTRACT
Tuberculous lymphadenitis is the most common cause of extra-pulmonary tuberculosis (TB)
in developing countries. Lymphadenitis caused by non-tuberculous mycobacteria (NTM)
requires consideration, particularly in immunocompromised patients and children in
developed countries. Fine Needle Aspiration Biopsy (FNAB) offers a valuable specimen
collection technique, but culture confirmation, mycobacterial speciation and drug resistance
testing (if indicated) is often unavailable in TB endemic areas and result in unacceptable
diagnostic delay.
We evaluated the diagnostic value of high-resolution DNA melting (HRM) analysis in
the diagnosis of mycobacterial lymphadenopathy using FNAB and an inexpensive transport
medium.
Specimens were collected from patients referred to the FNAB Clinic at Tygerberg
Hospital (June 2007-May 2008) with clinical mycobacterial lymphadenitis. Cytology, culture
and HRM were performed on all specimens. The reference standard for disease was defined
as positive cytology (morphological evidence plus mycobacterial visualization) and/or a
positive culture
90
Specimens were collected from 104 patients and mycobacterial disease was confirmed
in 54 (51.9 %); 46 M tuberculosis, 1 M. Bovis BCG, 1 NTM, and 8 not speciated. Cytology
was positive in 83.3 % (45 /54) and culture in 72.2 % (39/54) of patients. HRM identified
57.4 % (31/54) of cases. Using the defined reference standard we recorded 51.9 % sensitivity
and 94.0 % specificity (positive predictive value 90.3%) with HRM analysis.
HRM analysis allowed rapid and species specific diagnosis of mycobacterial lymph
adenitis in the majority of patients, permitting early institution of appropriate therapy.
Optimization of this technique requires further study.
INTRODUCTION
Resource limited countries carry the brunt of the global tuberculosis (TB) epidemic,
particularly those affected by the parallel human immunodeficiency virus (HIV) pandemic.1
According to the most recent World Health Organisation (WHO) estimates, in 2007 there
were 9.3 million incident (newly diagnosed) TB cases of whom 1.4 million (14.8%) were
HIV-infected. The African region accounted for 79% of HIV-infected TB cases.1 Although
TB incidence rates seem to plateau and/or decline in most regions, absolute numbers continue
to rise due to increases in population size. The projected scale of the epidemic and ongoing
transmission of drug resistant TB remains alarming.1 WHO estimated that between 2000 and
2010 1 billion people will be newly infected with M. tuberculosis, resulting in 200 million TB
cases and 35 million deaths.2
Peripheral lymphadenitis is the most common extra-pulmonary manifestation of TB.3-4
TB lymphadenitis is also the most common cause of persistent cervical lymphadenopathy in
children from TB endemic areas.5-7 In developed countries where the incidence of TB is low,
non-tuberculous mycobacteria (NTM) are frequently responsible for mycobacterial
lymphadenitis, particularly in children and HIV-infected immune compromised adults.3, 8-9 In
patients with a peripheral lymph node mass fine needle aspiration biopsy (FNAB) is a
91
valuable and underutilized specimen collection technique. This simple and safe procedure
allows rapid confirmation of mycobacterial disease using cytomorphology and direct
mycobacterial visualization with either Ziehl-Neelsen (ZN) staining or fluorescence
microscopy.
Mycobacterial culture is required for accurate species determination (speciation) and
drug susceptibility testing, 10-13 since organisms in the M. tuberculosis complex are
morphologically indistinguishable. The amount of material harvested during FNAB is
minimal, and the needle needs to be rinsed at the bedside in liquid medium to facilitate
culture. Although FNAB can be safely performed as an outpatient procedure by well trained
nurses 14, the need for direct inoculation and unavailability of liquid culture tubes limited
decentralization. Use of an inexpensive transport medium for direct inoculation has been
described,15 which should facilitate mycobacterial culture from FNAB’s performed in rural
clinics and hospitals. Direct bedside inoculation at the time of FNAB collection provides an
excellent diagnostic yield but culture results may take up to 6 weeks and requires additional
speciation.10, 16
Performing nucleic acid amplification tests (NAATs) on FNAB specimens may
provide a rapid species specific diagnosis and expedite access to appropriate therapy. A recent
systematic review demonstrated highly variable results with NAATs to diagnose TB
lymphadenitis, reported sensitivities ranged between 2-100% (specificities 28-100%).17 Most
NAATs analyze the polymerase chain reaction (PCR) products by gel electrophoresis or other
open tube formats, which afford the opportunity for laboratory cross contamination. These
technically challenging techniques will pose problems in countries with limited laboratory
resources. High resolution DNA melt analysis (HRM) is a simple “closed-tube” technique
that reduces the risk of cross contamination. Specific PCR amplicons are identified according
to their characteristic DNA melting profile. The amplicons are combined with a saturating
dye that fluoresces in the presence of double stranded DNA. It is heated through a range of
92
temperatures while fluorescence is monitored. 18 As the double stranded DNA dissociates
(melts) into single strands the fluorescence decreases. The melting peak of the specific
infectious agent is identified. The amplicon specific for M.tuberculosis melt at 90.5°C and
other members of the M.tuberculosis complex at 86°C. This simple technique can be used in
routine diagnostic laboratories.18-19
The aim of our study was to evaluate the value of PCR-based HRM analysis,18 to
provide a rapid and accurate diagnosis of mycobacterial disease using routinely collected
FNAB specimens directly inoculated into liquid transport medium.
MATERIALS AND METHODS
All adults and children referred to the FNAB Clinic at Tygerberg Hospital (June 2007-May
2008) with a superficial lymph node mass suggestive of possible mycobacterial lymphadenitis
and in whom written informed consent to participate in the study was obtained, were
included.
Specimen Collection
FNAB was done following standard protocol as an outpatient procedure by a pathologist.10
The lymph node was stabilized by the aspirator and two needle passes were performed using a
23g or 25g needle attached to a 10 ml syringe while applying a constant suction of no more
than 2 ml. From each aspirate two smears were prepared: one fixed with commercial cytology
fixative for Papanicolaou staining and the other air dried for the Giemsa and subsequent ZN
staining. The residual material in the syringe and needle was collected by withdrawing an
aliquot of liquid growth media from the TB transport bottle into the syringe and then
expelling the contents back into the bottle. No additional needle passes were performed to
collect material for microbiology or PCR. The TB transport bottles were prepared “in-house”
in a laminar flow cabinet: 10ml headspace glass vials containing 1 ml of Middlebrook 7H9
93
broth (with 0.2% glycerol and 0.05% Tween 80 added), sealed with 20mmTFE/Sil Septa and
20mm Aluminium open top seals and autoclaved at 120º C for 20min.20
Cytology
Stained smears were evaluated for adequacy and to make a morphological diagnosis. If any
specific lesion such as a lymphoepithelial cyst or neoplasia was present, this was reported.
Alternatively we reported “non-specific reactive lymphadenopathy” or “cytologically
consistent with mycobacterial infection”, based on specific morphological criteria.21-23 In
immunocompetent patients with TB, epithelioid granulomata and epithelioid histiocytes may
be identified in a background of reactive lymphocytes and plasma cells (Fig 1). A small
amount of amorphous necrosis may also be identified. In immunocompromised patients, the
cytological picture is that of abundant necrosis in which neutrophils and cellular debris are
prominent (Fig 2). In patients with lymphadenitis due to M bovis BCG or NTM infection,
histiocytes with abundant foamy cytoplasm may be present.8, 24-25
Mycobacteria were visualized using a ZN stain performed on a Giemsa stained slide
according to a slightly modified technique (Fig 3). Smears were differentiated in 3% acid
alcohol for one minute and counterstained with 1% methylene blue for 10 seconds only. In
addition, one of the Papanicolaou stained slides was screened using a Zeiss Axiophot
microscope with a fluorescent attachment and a wide-band blue excitation filter (450–480
nm); mycobacteria fluoresce as brilliant yellow bacilli, thin and slightly curved with polar
enhancement and a uniform length of 2.0 –2.7 microns (Fig 4).11
Microbiology
In the laboratory, 0.5ml of the media was aspirated from the TB transport bottle, inoculated
into a separate MGIT tube containing the PANTA antibiotic mixture and incubated in a
BACTEC MGIT 960 machine for 42 days. All positive MGIT tubes were confirmed to
contain acid fast bacilli in the absence of bacterial contamination by ZN staining.
Mycobacteria were identified as M. tuberculosis or other by standard PCR.26
94
Extraction of mycobacterial DNA
The TB transport bottles containing the remaining 0.5ml media were stored at -20°C until
further analysis. Mycobacteria within a 250µl aliquot were pelleted by centrifugation at full
speed (14 000rpm) for 10 minutes. The supernatant was discarded and the bacterial pellet was
resuspended in 1ml phosphate buffered saline and re-centrifuged at full speed (14 000rpm) to
remove residual blood. Thereafter the bacterial pellet was resuspended in 30µl ddH2O and
heat inactivated at 100°C for 20 minutes. The lysed bacterial extract was stored on filter paper
(FTA® Classic Card Collection, Storage and Purification system, Whatman, UK). Genomic
DNA was eluted from the filter paper cards following manufacturers’ instructions (high pH,
low pH protocol) and served as the template for subsequent PCR amplification.
Speciation and HRM Analysis
Primers were designed to amplify the Region of Deletion 9 (RD9), present in M. tuberculosis
and M. canetti, but absent from all other members of the M. tuberculosis complex. PCR was
performed as previously described 19 using an annealing temperature of 62°C and the
following primers RD9Fs1 5’-CAA GTT GCC GTT TCG AGC C-3’, RD9FR 5’-GCT ACC
CTC GAC CAA GTG TT-3’ and RD9INT 5’-CAA TGT TTG TTG CGC TGC-3’. 27
Resulting amplification products underwent HRM analysis in a RotorgeneTM6000 real-time
rotary analyzer (Corbett Life Science, Australia). The thermal denaturation profiles were
measured as previously described.19 The infectious agent was identified by the Rotorgene
TM
6000 software according to the presence of derivative melt peaks located within defined
temperature bins. M. tuberculosis PCR products melt at 90°C (M. tuberculosis could not be
differentiated from M. canettii, but M. canettii is very rarely observed and has not been
recorded in the study setting) while PCR products of other members of the M. tuberculosis
complex melted at 86°C. (Fig.4)
95
Reference standard for mycobacterial disease
The reference standard for the presence of mycobacterial disease17 was defined as positive
cytology (morphology consistent with mycobacterial infection plus direct visualization of the
organism) and/or positive culture with speciation.
Statistical Analysis
We assessed the diagnostic accuracy of PCR-based HRM analysis compared to cytology,
mycobacterial culture and the reference standard as defined using Pearson’s chi-square (X2)
and Fisher’s exact tests.28 All analyses were conducted using Statistica Version 8.29 Ethics
approval was obtained from the Institutional Review Board of Stellenbosch University
(N05/03/043).
RESULTS
FNAB specimens were collected from 104 patients with possible mycobacterial
lymphadenitis in whom complete cytopathology and mycobacterial culture results were
available together with PCR-based HRM analysis. The median age was 30 years with a range
from 4 months to 62 years. Children less than 13 years of age compromised 21.2 % of the
study population. There was no significant difference in gender distribution. HIV status was
known in 46.1% of patients (48 /104) and 75% of these were HIV positive (36/48). There
were 23 HIV-positive patients and 6 HIV negative patients with mycobacterial disease as
defined by the reference standard above. Table 1 summarizes the demographics and
diagnostic outcome of the study population.
Applying the defined reference standard 54 of the 104 patients (51.8 %) were
diagnosed with mycobacterial lymphadenitis. Mycobacterial culture was positive in 39/54
(72.2 %) patients; 37 M. tuberculosis, one M. bovis BCG and one NTM. The child with M.
bovis BCG was less than two years old; four others in this age group were positive for M.
tuberculosis. Cytology was positive in 83.3.9 % (45/54) of the cases. Cytology and culture
96
were both positive in 30/54 (55.6%) cases, while HRM analysis was positive in 57.4 %
(31/54) of cases identified using the reference standard. The calculated sensitivity of HRM
analysis was 51.9 % and the specificity 94.0 % (Fig 2) with a negative predictive value of
64.4 % and a positive predictive value of 90.3 % (Fig 5). Concordance between the different
diagnostic modalities and HRM analysis was poor, with Kappa values of 0.39 vs. cytology,
0.27 vs. mycobacterial culture, 0.19 vs. both cytology and culture, and 0.45 vs. cytology
and/or culture (the defined reference standard).
To investigate a possible association between mycobacterial load and the result of
PCR-based HRM analysis we tested for any correlation between the time to positivity (TTP,
often used as a surrogate of mycobacterial load) and PCR outcome. Among those with
positive mycobacterial cultures no correlation between the time to positivity and PCR-based
HRM analysis could be demonstrated; mean TTP in PCR positive cases was 18 days
compared to 22 days in PCR negative cases (p=0.22).( Fig 5)
DISCUSSION
With global coordination of control efforts TB incidence rates seem to be stabilizing and
showing signs of decline in all six world regions. However, the TB disease burden in areas
plagued by the concurrent HIV pandemic remains at unprecedented levels.1 Although there is
some assistance for the development of strong laboratory networks in developing countries,
the existing infrastructure remains poor in most TB endemic areas. South Africa, despite its
relatively good infrastructure and health care services, in 2007 reported a national TB
incidence of 739.6/100,000 population and an adult TB incidence in the Western Cape
Province, a high burden region, of 1005.7 /100.000.30
In TB endemic areas with ongoing transmission, children constitute a significant
percentage of the total caseload, estimated at 15-20%.31 Immune immaturity and/or
compromise, most often due to HIV infection,32-33 influence the risk to develop extra
97
pulmonary and/or disseminated TB as well as the likelihood of rapid disease progression and
TB-related mortality.34-35 Autopsy studies in Africa 33, 36-39 have shown that up to 54% of
deaths in HIV-infected adults and 20% in HIV-infected children are due to TB. Reducing this
mortality and morbidity necessitates early detection, efficient diagnosis and timely institution
of appropriate therapy.
FNAB is a simple, non-invasive specimen collection technique in patients with
peripheral lymph node masses 10. It assists with rapid diagnosis of mycobacterial disease, but
also helps to rule out alternative diagnoses that may require urgent treatment such as
neoplasia. Diagnostic sensitivity is dependent on the experience of the pathologist and for
mycobacterial disease varies from 32-78%.10, 40-41 Identification of the organism is essential
particularly in immune compromised patients, as other opportunistic infections such as fungal
organisms may elicit a similar morphological reaction pattern. Mycobacteria may be
visualized using ZN staining, which is a very simple stain but has suboptimal sensitivity (2062%).10, 17, 42-43 Autofluorescence, using the ability of mycobacteria on Papanicolaou stains
smears to fluoresce using a blue excitation filter requires no additional staining and improves
sensitivity (65-67%).10-11 Culture is often regarded as the “gold standard”, but is limited by
poor sensitivity (reported to be as low as 2-34%) 41, 44 and is highly dependent on the culture
medium and inoculation technique used. A positive result is delayed by 1-6 weeks and
requires additional PCR-based testing for speciation.
Direct NAAT application provides results in 3 to 6 hours and has been evaluated with
respiratory and non-respiratory specimens such as FNAB’s.17, 43-47 Test sensitivity with
respiratory specimens is highly dependent on mycobacterial load and use current PCR-based
tests are only advised in patients with sputum smear-positive TB. Results with extrapulmonary disease have been variable but recent studies have shown excellent results. Use of
a nested PCR in cervical TB lymphadenitis in Mexican children showed a sensitivity of 96%
and a specificity of 93%,47 whilst a study utilizing DNA from dried and fixed cytology smears
98
showed a sensitivity of 85% and specificity of 95% using nested PCR, although reference
standards are not necessarily based on bacteriological confirmation of disease.48
The majority of these studies use agarose gel electrophoresis to visualize the products, which
is not practical in a busy microbiology laboratory. The open tube nature of the procedure also
allows the release of amplicons, which pose a real risk of cross-contamination. A recent
systematic review evaluating the diagnostic utility of NAATs in TB lymphadenitis identified
36 peer reviewed publications.17 The authors found marked variation in populations, test
techniques, reference standards, volume of material utilized and quality indicators. Few
studies controlled for cross contamination and inhibitors in clinical samples, thereby
increasing the possibility of both false positive and false negative results. The reference
standard used most frequently was culture, although the authors regard the sensitivity of
culture from FNAB to be about 62%. Using an imperfect reference standard may lead to an
underestimation of test performance and this is a major limitation that is rarely
acknowledged.49
The current study attempted to address some of the shortcomings noted above. The
reference standard used is a combination of positive cytology, defined as cytomorphological
features consistent with mycobacterial infection combined with identification of the organism
using ZN staining or autofluorescence and/or positive culture. The residual material from the
fine needle aspirates was collected in TB transport medium in a sealed bottle, minimizing the
possibility of contamination. After washing and concentration of the specimen the pellets
were placed on filter paper for storage and to remove inhibitors present in the specimen.
After extraction of DNA, the products were amplified using primers designed to amplify the
Region of Deletion 9 (RD9), present in M. tuberculosis and M. canetti, but absent from all
other members of the complex.19 The amplified products were then identified using high
resolution melt analysis, which is a closed tube format that minimizes the possibility of crosscontamination.
99
This technique is rapid and simple, and the equipment required for the HRM such as
the Rotorgene 6000TM real-time rotary analyzer (Corbett Life Science, Australia) is relatively
inexpensive. Up to 72 samples may be processed in a cycle which takes approximately 3
hours.
The results of PCR using this technique did not differ significantly from that of
cytology or culture when these were assessed independently against the reference standard.
The sensitivity is relatively low, but the specificity is high at 94% as is the PPV of 90%. This
enables appropriate therapy to be implemented early with a high degree of confidence in the
majority of patients with disease, while continuing with culture in the PCR negative cases.
Limited sample volume may have accounted for the low sensitivity observed. No
additional needle passes or aspirates were performed and the material collected was limited to
the residual amount left in the needle. Sensitivity may be improved if an additional needle
pass is performed to obtain material for PCR. No single diagnostic modality was adequately
sensitive to enable it to be used alone. FNAB provides material for cytology, culture and PCR
using the transport medium described. This is of particular value in resource limited countries
where laboratories tend to be centralized.1 Cytology slides once prepared are stable at room
temperature, as is the transport medium where the mycobacterial organism has been shown to
remain viable for up to 7 days at room temperature, enabling both culture and PCR. 20
Ideally, the PCR technique could be refined, increasing the sensitivity and thereby
eliminating the need for culture, which is costly and lengthy. Use of the transport bottle
facilitates both culture and PCR analysis, but material may be deposited directly onto filter
paper for subsequent PCR analysis. Refined techniques need to remain simple inexpensive
and appropriate for use in routine laboratories in countries with limited resources and skills
that bear the burden of this devastating and persistent disease.
100
ACKNOWLEDGEMENTS
Financial support was received from the NHLS Trust Fund, National Health Laboratory
Services, South Africa and The Division of Molecular Biology and Human Genetics,
Stellenbosch University.
We thank Prof Martin Kidd (Stellenbosch University) for statistical assistance
No industry funding was received to perform this study, and none of the authors have any
commercial interest to declare.
This study was done in partial fulfillment of a PhD thesis (CA Wright, KGP Hoek)
REFERENCES
[1]
World Health Organization. Global Tuberculosis Control: Surveillance, Planning,
Financing: WHO Report 2007. Geneva: World Health Organization; 2007. Report No.:
WHO/HTM/TB/2007.376.
[2]
Shingadia D, Novelli V. Diagnosis and treatment of tuberculosis in children. Lancet
Infect Dis 2003;3:624-632.
[3]
Golden MP, Vikram HR. Extrapulmonary tuberculosis: an overview. Am Fam
Physician 2005;72:1761-1768.
[4]
Marais BJ, Wright CA, Schaaf HS, et al. Tuberculous Lymphadenitis as a Cause of
Persistent Cervical Lymphadenopathy in Children From a Tuberculosis-Endemic Area.
Pediatr Infect Dis J 2006;25:142-146.
[5]
Marais BJ, Gie RP, Schaaf HS, et al. The natural history of childhood intra-thoracic
tuberculosis: a critical review of literature from the pre-chemotherapy era State of the Art. Int
J Tuberc Lung Dis 2004;8:392-402.
[6]
Marais BJ, Gie RP, Schaaf HS, et al. The clinical epidemiology of childhood
pulmonary tuberculosis: a critical reading of literature from the pre-chemotherapy era. Int J
Tuberc Lung Dis 2004;8:278-285.
101
[7]
Moore SW, Schneider JW, Schaaf HS. Diagnostic aspects of cervical
lymphadenopathy in children in the developing world: a study of 1,877 surgical specimens.
Pediatr Surg Int 2003;19:240-244.
[8]
Hazra R, Robson CD, Perez-Atayde AR, Husson RN. Lymphadenitis Due to
Nontuberculous Mycobacteria in Children: Presentation and Response to Therapy. Clin Infect
Dis 1999;28:123-129.
[9]
Flint D, Mahadevan M, Barber C, Grayson D, Small R. Cervical lymphadenitis due to
non-tuberculous mycobacteria: surgical treatment and review. Int J Pediatr Otorhinolaryngol
2000;53:187-194.
[10]
Wright CA, van der Burg M, Geiger D, et al. Diagnosing Mycobacterial lymphadenitis
in children using Fine Needle Aspiration Biopsy: Cytomorphology, ZN staining and
Autofluorescence – making more of less. Diagn Cytopathol 2008;36:245-251.
[11]
Wright CA, van Zyl Y, Burgess SM, Blumberg L, Leiman G. Auto fluorescence of
Mycobacteria on Lymph Node Aspirates – A glimmer in the dark? Diagn Cytopathol
2004;30:257-260.
[12]
Thomas JO, Adeyi D, Amanguno H. Fine Needle Aspiration in the Management of
Peripheral Lymphadenopathy in a Developing Country. Diagn Cytopathol 1999;21:159-162.
[13]
Bezabih M, Marium DW, Selassie SG. Fine needle aspiration cytology of suspected
tuberculous lymphadenitis. Cytopathology 2002;13:284-290.
[14]
Leiman G. The Registered Nurse as Aspirator:Experience in a Busy Cytopathology
Unit. Acta Cytol 1995;39:1010.
[15]
Wright CA, Bamford C, Prince Y, et al. Mycobacterial Transport medium for routine
culture of Fine Needle Aspiration Biopsies. Arch Dis Child 2009;In press.
[16]
Wright CA, Hesseling A, Bamford C, et al. Fine-needle aspiration biopsy: a first-line
diagnostic procedure in paediatric tuberculosis suspects with peripheral lymphadenopathy. Int
J Tuberc Lung Dis 2009;in press.
102
[17]
Daley P, Thomas S, Pai M. Nucleic acid amplification tests for the diagnosis of
tuberculous lymphadenitis: a systematic review. Int J Tuberc Lung Dis 2007;11:1166-1176.
[18]
Reed GH, Kent JO, Wittwer CT. High-resolution DNA melting analysis for simple
and efficient molecular diagnostics. Pharmacogenomics 2007;8:597-608.
[19]
Hoek KGP, Gey van Pittius NC, Moolman-Smook H, et al. Fluorometric Assay for
Testing Rifampin Susceptibility of Mycobacterium tuberculosis Complex. J Clin Microbiol
2008;46:1369-1373.
[20]
Vermaak A, Wright CA, Bamford C, et al. TB Transport Medium for Fine Needle
Aspiration Biopsies TB Conference 2008. Durban, South Africa 2008.
[21]
Purohit SD, Purohit V, Mathur ML. A Clinical Scoring System as Useful as FNAC in
the Diagnosis of Tuberculous Lymphadenitis in HIV Positive Patients. Current HIV Research
2006;4:459-462.
[22]
Kocjan G, Miller R. The cytology of HIV-induced immunosuppression. Changing
pattern of disease in the era of highly active antiretroviral therapy. Cytopathology
2001;12:281-296.
[23]
Nayak S, Mani R, Kavatkar AN, Puranik SC, Holla VV. Fine-needle aspiration
cytology in lymphadenopathy of HIV-positive patients. Diagn Cytopathol 2003;29:146-148.
[24]
Schaaf HS, Zumla A. TUBERCULOSIS: A Comprehensive Clinical Reference.
Philadelphia: SAUNDERS, Elsevier 2009.
[25]
Al-Bhlal LA. Fine-needle aspiration cytology of postvaccinial disseminated bacillus
Calmette-Guerin infection. Diagn Cytopathol 2001;24:333-335.
[26]
De Wit D, Steyn L, Shoemaker S, Sogin M. Direct detection of Mycobacterium
tuberculosis in clinical specimens by DNA amplification. J Clin Microbiol 1990;28:24372441.
103
[27]
Warren RM, Gey PNC, Barnard M, et al. Differentiation of Mycobacterium
tuberculosis complex by PCR amplification of genomic regions of difference. Int J Tuberc
Lung Dis 2006;10:818-822.
[28]
Agriesti A. Categorical Data Analysis. 2nd ed. Hoboken, New Jersey: Wiley 2002.
[29]
StatSoft Inc. Statistica, Version 8. Tulsa, Oklahoma: Statsoft, Inc. 2007.
[30]
Health Systems Trust. 2007 6 March 2009 [cited 6 March 2009]; Available from:
http://www.hst.org.za/healthstats/16/data
[31]
Marais BJ, Gie RP, Schaaf HS, et al. Childhood pulmonary tuberculosis: old wisdom
and new challenges. Am J Respir Crit Care Med 2006;173:1078.
[32]
Braun MM, Cauthen G. Relationship of the human immunodeficiency virus epidemic
to paediatric tuberculosis and Bacillus Calmette-Guerin vaccination. Pediatr Infect Dis J
1992;11:220-227.
[33]
Madhi SA, Huebner RE, Doedens L, et al. HIV-1 coinfection in children hospitalised
with tuberculosis in South Africa. Int J Tuberc Lung Dis 2000;4:448-454.
[34]
Marais BJ, Graham SM, Cotton MF, Beyers N. Diagnostic and management
challenges for childhood tuberculosis in the era of HIV. J Infect Dis 2007;196:76-85.
[35]
Walters E, Cotton MF, Rabie H, et al. Clinical presentation and outcome of
Tuberculosis in Human Immunodeficiency Virus infected children on anti-retroviral therapy.
BMC Pediatrics 2008;8:1.
[36]
Lucas SB, Hounnou A, Peacock C, et al. The mortality and pathology of HIV
infection in a West African city. AIDS 1993;7:1569-1579.
[37]
Chintu C, Mudenda V, Lucas SB. Lung diseases at necropsy in African children dying
from respiratory illnesses; a descriptive necropsy study. Lancet 2002;360:985-990.
[38]
Kiwanuka J, Graham SM, Coulter JBS, et al. Diagnosis of pulmonary tuberculosis in
children in an HIV-endemic area, Malawi. Ann Trop Paeds 2001;21:5-14.
104
[39]
Jeena PM, Pillay P, Pillay T, Coovadia HM. Impact of HIV-one coinfection on
presentation and hospital-related mortality in children with culture proven pulmonary
tuberculosis in Durban, South Africa. Int J Tuberc Lung Dis 2002;6:672-678.
[40]
Lau SK, Wei WI, Hsu C, Engzell UCG. Efficacy of fine needle aspiration cytology in
the diagnosis of tuberculous cervical lymphadenopathy. J Laryngol Otol 1990;104:24-27.
[41]
Singh KK, 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.
[42]
Gupta AK, Nayar M, Chandra M. Critical appraisal of fine needle aspiration cytology
in tuberculous lymphadenopathy. Acta Cytol 1992;36:391-394.
[43]
Gong G, Lee H, Kang GH, et al. Nested PCR for diagnosis of tuberculous
lymphadenitis and PCR-SSCP for identification of rifampicin resistance in fine-needle
aspirates. Diagn Cytopathol 2002;26:228-231.
[44]
van Coppenraet ESB, Lindeboom JA, Prins JM, et al. Real-Time PCR Assay Using
Fine-Needle Aspirates and Tissue Biopsy Specimens for Rapid Diagnosis of Mycobacterial
Lymphadenitis in Children. J Clin Microbiol 2004;42:2644-2650.
[45]
Ling DI, Flores LL, Riley L, Pai M. Commercial Nucleic-Acid Amplification Tests
for diagnosis of Pulmonary Tuberculosis in Respiratory Specimens: Meta-Analysis and MetaRegression. PLos ONE 2008:e1536.
[46]
Aljafari AS, Khalil EAG, Elsiddig KE, et al. Diagnosis of tuberculous lymphadenitis
by FNAC, microbiological methods and PCR: a comparative study. Cytopathology
2004;15:44-48.
[47]
Portillo-Gomez L, Murillo-Neri MV, Gaitan-Mesa J, Sosa-Iglesias EG. Nested
polymerase chain reaction in the diagnosis of cervical tuberculous lymphadenitis in Mexican
children. Int J Tuberc Lung Dis 2008;12:1313-1319.
105
[48]
Purohit MRMD, Mustafa TMP, Sviland LMDMP. Detection of Mycobacterium
tuberculosis by Polymerase Chain Reaction With DNA Eluted From Aspirate Smears of
Tuberculous Lymphadenitis. Diagnostic Molecular Pathology 2008;17:174-178.
[49]
Pai M, O'Brien R. Tuberculosis diagnostics trials:do they lack methodological rigor?
Expert Rev Mol Diagn 2006;6:509-514.
106
Table 1
Demographics and diagnostic outcome
Number of cases
Age
< 2 years
<13 years
> 13 years
Unknown
Gender
Male
Female
Unknown
HIV status
Positive
Negative
Unknown
Cytodiagnoses
Malignancy
Reactive node
Other (e.g. cyst)
Cytology positive mycobacterial infection
Cases with mycobacterial disease
Culture and/or cytology positive (Reference Std)
Cytology positive
Culture positive
Culture positive and cytology positive
PCR positive mycobacterial infection
Numbers
104
Percentage
6
22
77
5
5.8
21.2
74
5
48
51
5
46.2.
49.0
4.8
36
12
56
34.6
11.5
53.9
12
43
4
45
11.5
41.4
3.9
43.3
54
45
39
30
31
100
83.3
72.2
55.6
57.4
107
Figure 1. Epithelioid granuloma in an immune competent patient. (Papanicolaou, x400)
Figure 2. Abundant necrosis in which neutrophils and cellular debris are prominent
consistent with tuberculous lymphadenitis in an immune compromised patient.
(Papanicolaou, x400)
108
Figure 3. Modified ZN stain in a lymph node aspirate. (Ziehl-Neelson, x400)
Figure 4. Autofluorescence of mycobacteria in a lymph node aspirate.
(Papanicolaou, x1000 with a wide-band blue excitation filter)
109
Fig 5 Rotorgene software depicting the derivative melt peaks located within defined
temperature bins. M.tuberculosis products melted at 90°C and M.tuberculosis complex at
86°C.
110
CHAPTER 7
CONCLUSION
The aim of this dissertation was to demonstrate the value of Fine Needle Aspiration Biopsy in
the diagnosis of mycobacterial lymphadenitis with particular reference to children.
Mycobacterial disease in children poses a particular diagnostic and management
problem. It is a major cause of childhood morbidity and mortality in developing countries,
where it is under reported due to problems in obtaining an accurate diagnosis. The traditional
methods used to diagnose tuberculosis (TB) in adults are not feasible in children under the age
of 7-8 years, young children experience the greater disease burden and a significant
percentage (~10%) of children have extra pulmonary disease manifestations only.
Clinical criteria and chest radiography have diagnostic limitations which are compounded in
children who are HIV infected. In young children and in those with HIV/AIDS, early
diagnosis is essential, as dissemination may occur rapidly due to their immature or
compromised immune system.
We first established that TB lymphadenitis is a common clinical problem in children
in TB endemic areas and that FNAB is an efficient, simple and effective diagnostic modality
in determining the cause of lymphadenopathy in children.
We showed that FNAB provides superior diagnostic yields and a significant reduction
in time to diagnosis when compared with conventional laboratory specimens in children.
Autofluorescence was demonstrated to be an accurate, rapid and inexpensive method
for demonstrating mycobacteria on routinely stained cytology smears and this was confirmed
in a prospective study in children where FNAB was shown to provide a rapid and definitive
diagnosis in the majority of cases with TB lymphadenitis.
111
We developed a new reference standard for mycobacterial disease using FNAB positive cytology (cytomorphological evidence of disease and identification of the organism)
and/or positive culture.
Lymphadenopathy induced by BCG vaccination needs to be distinguished from TB
lymphadenitis. Although not central to this thesis, a small study during this time period
demonstrated cytomorphological features on FNAB which may assist in distinguishing
between these two organisms, together with clinical features. (See Appendix A)
Although cytomorphology may confirm mycobacterial disease, organisms in the M
tuberculosis complex are morphologically indistinguishable. This requires submission of
material for culture and subsequent speciation or submission of material for direct PCR.
As FNAB yields a miniscule amount of material, immediate bedside inoculation is required
for culture. We developed an inexpensive liquid transport medium which facilitated culture
from FNAB specimens, in which organisms remain viable for up to 7 days at room
temperature. This is particularly useful in resource limited countries where laboratories are
centralised and transport of specimens may be problematical.
Culture for speciation of mycobacterial organisms takes up to 6 weeks and this may
result in inappropriate treatment or delayed commencement of therapy. Using the transport
medium developed we established a rapid simple and inexpensive technique using PCR and
High Resolution Melt Analysis, which is feasible for implementation in routine diagnostic
laboratories.
In summary, the outcome of these studies has demonstrated the need for, and value of,
FNAB in the diagnosis of mycobacterial lymphadenitis particularly in children in TB endemic
areas.
We have improved the ability of cytology to confirm mycobacterial disease using
autofluorescence. We have developed a means to obtain a culture from FNAB using a novel
112
transport medium and have developed a simple nucleic acid amplification technique which
will enable rapid confirmation and speciation of mycobacterial disease.
There are many challenges that remain, but we are working to develop an even simpler
technique for collecting material from FNAB for PCR/HRM analysis which would provide
higher sensitivity by removing inhibitors in the sample.
Conventional fluorescent microscopy using mercury vapour lamps is expensive and
therefore not feasible in resource-limited countries, and we intend to demonstrate the utility of
inexpensive rewindable LED light sources in autofluorescence and auramine stained FNAB
smears. A retrospective comparative study has already been completed. (see addendum A)
The quest to find and establish the ideal modality for diagnosing mycobacterial
disease has been a long one for many researchers. The many clinical manifestations of this
elusive disease and the reality that it is a disease that affects the poor and less influential
segments of the global population compounds the problem.
Currently a lot of attention is focused on identifying improved methods for TB
diagnosis. The field is ever-changing and no sooner has a technique been optimised when it is
evident that a better methodology has emerged. We hope that this research has contributed a
small step towards success in the fight against this devastating disease.
113
REFLECTIVE ASSESSMENT OF CONTRIBUTION.
In summary this thesis makes the following theoretical, empirical and methodological
contributions to the literature.
Theoretical contribution:
FNAB is utilised in adult patients in neoplastic and infectious diseases, and to a lesser extent
in paediatric oncology, but it potential contribution in paediatrics is not sufficiently
appreciated. These results demonstrate the utility of FNAB in the diagnosis of TB, especially
in children. The studies show that FNAB is equal to or superior to conventional diagnostic
specimen types in children and that it is accurate, easy to perform, and yields rapid diagnostic
information.
Empirical contribution:
Countries with low economic resources and high healthcare resource constraints bear the
brunt of TB and HIV infections. This body of research demonstrates that FNAB can be
implemented in these settings in the diagnosis of mycobacterial disease in children. It shows
how FNAB and a novel transport medium solve problems in the collection of material for
culture and NAAT, and autofluorescence assists in the cytological diagnosis of mycobacterial
disease.
Methodological contribution:
The methodologies and concepts used in this thesis have been developed specifically for
utilisation in resource limited countries.
A new method of NAAT on FNAB material is developed and its benefits assessed.
This new method is appropriate for use in resource limited countries. Although NAAT’s have
114
been used in the diagnosis of mycobacterial disease on FNAB, the results are extremely
variable and impractical for implementation in routine diagnostic laboratories. The new
method utilises high resolution DNA melt analysis was developed which is rapid inexpensive,
minimises contamination and is applicable for implementation in routine diagnostic
laboratories. This was performed on specimens collected in the mycobacterial transport
medium.
A novel transport medium was developed to facilitate mycobacterial culture from
FNAB specimens, particularly in rural hospitals and clinics. Autofluorescence is used in the
research as an additional inexpensive and rapid means of bacteriological confirmation of
mycobacterial disease on cytology smears, which was not previously published. This was then
applied to fine needle aspirates in children with mycobacterial disease.
115
“I am not discouraged, because every wrong attempt discarded is another step forward”
Thomas A. Edison
I hope that this body of research will help others remember the inestimable honour we all
have in serving humankind by undertaking research into the diseases plaguing our nation. All
of us have an invaluable and unique role to play from the clinicians in their day-to-day contact
with patients experiencing the frustration of practical problems associated with diagnosis and
management of disease, to the biomedical scientists with out whose expertise solutions would
not be possible. It is through collaboration and commitment to a common goal that we will
achieve the Millennium Developmental Goal 6 of reducing the prevalence and death rate from
tuberculosis by 2015 through early diagnoses of new cases.
The primary, the most urgent requirement is the promotion of education. It is inconceivable
that any nation should achieve prosperity and success unless this paramount, this
fundamental concern is carried forward. The principal reason for the decline and fall of
peoples is ignorance. Today the mass of the people are uninformed even as to ordinary
affairs, how much less do they grasp the core of the important problems and complex needs of
the time.
It is therefore urgent that beneficial articles and books be written, clearly and definitely
establishing what the present-day requirements of the people are, and what will conduce to
the happiness and advancement of society. These should be published and spread throughout
the nation, so that at least the leaders among the people should become, to some degree,
awakened, and arise to exert themselves along those lines which will lead to their abiding
honor. The publication of high thoughts is the dynamic power in the arteries of life; it is the
very soul of the world. Thoughts are a boundless sea, and the effects and varying conditions
116
of existence are as the separate forms and individual limits of the waves; not until the sea
boils up will the waves rise and scatter their pearls of knowledge on the shore of life.
(Abdu'l-Baha, The Secret of Divine Civilization, p. 108)