INT J TUBERC LUNG DIS 12(4):417–423
© 2008 The Union
High level of discordant IGRA results in HIV-infected
adults and children
A. M. Mandalakas,* A. C. Hesseling,† N. N. Chegou,‡ H. L. Kirchner,§ X. Zhu,* B. J. Marais,† G. F. Black,‡
N. Beyers,† G. Walzl‡
*Department of Pediatrics, School of Medicine, Case Western Reserve University, Cleveland, Ohio, USA; †Desmond
Tutu TB Centre, Department of Pediatrics and Child Health, Faculty of Health Sciences, Stellenbosch University,
Tygerberg, ‡Division of Molecular Biology and Human Genetics, Department of Science and Technology/National
Research Foundation Centre of Excellence in Biomedical Tuberculosis Research and MRC Centre for Molecular and
Cellular Biology, Faculty of Health Sciences, Stellenbosch University, Tygerberg, South Africa; §Center for Health
Research, Geisinger Health System, Danville, Pennsylvania, USA
S U M M A R Y
SETTING: Tygerberg district, Western Cape Province,
OBJECTIVE: To measure the agreement of two interferon-
gamma release assays (IGRAs) and the tuberculin skin
test (TST) for the detection of Mycobacterium tubercu-
losis infection in human immunodeficiency virus (HIV)
infected adults and children in a setting highly endemic
for tuberculosis (TB).
DESIGN: Cross-sectional study.
RESULTS: In HIV-infected adults (n ? 20) and children
(n ? 23), tests yielded discordant results, with 61% of
individuals testing positive with T-SPOT.TB, 41% with
TST and 28% with QuantiFERON® TB Gold (QTF). In
children, there was poor agreement between the TST and
T-SPOT.TB (kappa [?] ? ?0.02), but moderate agree-
ment between the TST and QTF (? ? 0.44). In adults,
there was moderate agreement between the TST and
T-SPOT.TB (? ? 0.43), and the TST and QTF (? ?
0.46). In children and adults, there was fair agreement
between the T-SPOT.TB and QTF (? ? 0.33). Twenty
per cent of adults had ?1 indeterminate IGRA results.
CONCLUSIONS: There is poor to moderate agreement
between the TST and IGRAs in HIV-infected adults and
children. T-SPOT.TB may have improved sensitivity
for detection of M. tuberculosis infection in HIV-
infected individuals compared to the QTF and the TST.
In HIV-infected individuals, IGRA test properties are af-
fected by test cut-off point and nil control responses.
KEY WORDS: interferon-gamma release assays; ESAT-6;
CFP-10; M. tuberculosis; exposure age; HIV; CD4
NINE MILLION people contract tuberculosis (TB)
and 2 million people die from TB annually.1 Children
represent 11–15% of this disease burden.2 The risk of
progression to active TB following Mycobacterium
tuberculosis infection in human immunodeficiency virus
(HIV) infected adults is 5–10% per year,3 and persists
despite antiretroviral treatment (ART).4 The risk of pro-
gression to disease following M. tuberculosis infection
in HIV-infected children is at least 6–8 fold higher than
in non-HIV-infected children,5 with a TB incidence of
9.2% recently recorded in HIV-infected children who
did not receive isoniazid preventive treatment (IPT).6
The diagnosis of M. tuberculosis infection is com-
plicated by the lack of a gold standard. The negative
predictive value (NPV) of the tuberculin skin test (TST)
is reduced in immunocompromised individuals, the very
young and/or severely malnourished children and in
the presence of severe TB.7 TST sensitivity in detect-
ing recent infection may be limited, as demonstrated
by individuals who convert their TST 6–12 weeks after
documented M. tuberculosis exposure.8 The positive
predictive value (PPV) of TST is reduced in popula-
tions with exposure to non-tuberculous mycobacteria
(NTM) and/or M. bovis bacille Calmette-Guérin (BCG)
vaccination.7 The potential benefits of an accurate
diagnostic test of M. tuberculosis infection in HIV-
infected individuals include timely institution of pre-
ventive or curative treatment.
T-cell-based interferon-gamma (IFN-?) release as-
says (IGRAs), utilising early secretory antigenic target
6 (ESAT-6), culture filtrate protein 10 kDa (CFP-10)
and TB7.7, may offer enhanced sensitivity in the diag-
nosis of M. tuberculosis infection in HIV-infected
individuals compared to the TST.9–11 However, few
studies have directly compared IGRAs and the TST in
HIV-infected adults and children in TB-endemic set-
tings. Discordant results between TST and IGRAs
have not been well studied in HIV-infected individuals.
Correspondence to: A M Mandalakas, Division of Global Child Health, Department of Pediatrics, School of Medicine, Case
Western Reserve University, 11100 Euclid Avenue—MS 7052, Cleveland, OH 44106, USA. Tel: (?1) 216 844 3224. Fax:
(?1) 216 844 6265. e-mail: email@example.com
Article submitted 17 July 2007. Final version accepted 13 December 2007.
The International Journal of Tuberculosis and Lung Disease
We conducted a cross-sectional study to assess the
agreement of two commercially available IGRAs in
relation to the TST and investigated the impact of
age, M. tuberculosis exposure and CD4 count on test
responses for the detection of M. tuberculosis infec-
tion in HIV-infected children and adults.
Study setting and participants
The present study was conducted in the Western Cape
Province, South Africa, where the TB notification rate
was 841 per 100000 in 200212 and the incidence of
TB in children was 407/100000 in 2004.13 The BCG
(Danish strain, 1331, Statens Serum Institute, Copen-
hagen, Denmark) vaccination is routinely adminis-
tered at birth. HIV prevalence among women attend-
ing public antenatal facilities was 15.7% in 2005.14
From September 2005 to March 2006, HIV-infected
adults and children routinely attending the Tygerberg
Academic Hospital Family HIV clinic were consecu-
tively recruited. Individuals were excluded if they were
acutely ill, had active TB or were receiving IPT.
Informed consent was provided. Institutional ap-
proval was obtained from Stellenbosch University and
Tygerberg Academic Hospital. All children in house-
hold contact with a TB index case were referred for IPT.
Data were collected on age, BCG vaccination status,
history of household contact with a bacteriologically
confirmed adult pulmonary TB case within the pre-
ceding 3 months, ART and CD4 count. Once a 10 ml
blood sample had been obtained, a TST using two tu-
berculin units (TU) of RT 23 purified protein deriva-
tive (PPD) (Mantoux PPD, Statens Serum Institute)
was placed on the volar aspect of the left forearm and
read within 48–72 h by a single study nurse, follow-
ing standardisation, using the ball-point pen and
ruler method.15,16 A positive TST was defined as an
induration of ?5 mm.
Blood was transported at room temperature and pro-
cessed within 3 h. Specimens with inadequate blood
volumes, delayed transport time or clotting were doc-
umented as ‘failed phlebotomy’. If blood volumes were
inadequate, the T-SPOT.TB test was preferentially
completed. Laboratory analyses were blinded.
The T-SPOT.TB and QuantiFERON® TB Gold
(QFT) tests were performed according to the manu-
facturers’ instructions (Cellestis, Carnegie, VIC, Aus-
tralia, and Oxford Immunotec, Abingdon, UK). For
the T-SPOT.TB test, if the negative (nil) control had
0–5 spots, a result was considered positive if (Panel A
or Panel B spot count) ? (nil control spot count) was
?6; if the nil control had ?6 spots, a result was con-
sidered positive if (Panel A or Panel B spot count) ?2 ?
(nil control spot count), irrespective of the mitogen
(positive control) spot count. A positive control was
defined as ?20 spots. Indeterminate T-SPOT.TB results
were defined as a mitogen spot count ?20 in conjunc-
tion with negative antigen panel responses.
For the QTF, a positive result was defined as ESAT-6
and/or CFP-10 ? nil control ? 0.35 international units
(IU)/ml regardless of mitogen value, and a negative
result as ESAT-6 and/or CFP-10 ? nil control ? 0.35
IU/ml and Mitogen ? nil ? 0.5 IU/ml. Indeterminate
QTF results were defined as ESAT-6 and/or CFP-10 ?
nil ? 0.35 IU/ml and mitogen ? nil ? 0.5 IU/ml.
T-SPOT.TB results were read using an automated
spot counter (Zeiss, Gottingen, Germany). HIV sta-
tus was confirmed by enzyme-linked immunosorbent
assay (ELISA) in individuals aged ?18 months or by
polymerase chain reaction (PCR) in children aged
?18 months. On-site proficiency testing and quality
assurance measures were completed for both IGRAs by
the manufacturers (Oxford Immunotec and Cellestis).
Data are expressed as frequencies and percentages for
categorical means and standardisations for continu-
ous variables. Children and adults were compared
using the Pearson’s ?2 and Wilcoxon rank-sum tests.
Rates of positive results were compared across tests
using the McNemar ?2 test for correlated proportions.
Agreement between tests was estimated using Co-
hen’s kappa (?) coefficient. The association between
each diagnostic test and reported M. tuberculosis ex-
posure was estimated using logistic and linear regres-
sion models. For linear regression, continuous IGRA
responses were standardised to achieve a distribution
that approximated normality with a mean of zero and
a standard deviation of one. Regression models con-
trolled for the potential confounding effect of CD4
Forty-three HIV-infected individuals (23 children and
20 adults) were enrolled (Table 1). Forty-eight per cent
of the children, but none of the adults, were on ART.
The mean CD4 count in children was 1162.3 (normal
range 1000–1800), whereas in adults it was 334.3 (nor-
mal range 500–2010). There was no significant differ-
ence between adults and children with regard to the pro-
portion BCG-vaccinated or with TB contact.
Test results were positive in 61% for the T-SPOT.TB
test, 41% for the TST and 28% for QTF (Table 1).
Due to failed phlebotomy, only 12 children had QTF
results. The proportion of children with a positive
T-SPOT.TB did not differ between children who did
and did not complete QTF testing. Four adults did not
have TST results recorded. Median TST values in
IGRA discordance in HIV-infected individuals
children and adults with a TST ?0 mm were respec-
tively 11 mm and 17 mm. The TST was positive in more
adults than children (62.5% vs. 26.1%, P ? 0.02). In
contrast, there was no significant difference in the
proportion of positive T-SPOT.TB or QTF results in
children compared to adults. QTF and T-SPOT.TB
produced indeterminate results in respectively 15%
(3/20) and 10% (2/20) of adults. No indeterminate re-
sults were observed in children.
Regardless of age group, subjects were more likely to
have a positive T-SPOT.TB test than QTF (P ? 0.05)
(Table 1). TST was also more often positive than the
QTF in all subjects and in the adult group (P ? 0.05),
but not in children. There was no significant difference
between the proportion of positive TST and T-SPOT.TB
results in any group. All adults with indeterminate
IGRA results had low CD4 counts (mean count ? 209)
and TST values of zero. There was moderate agreement
between the TST and T-SPOT.TB in adults (? ? 0.43),
and the TST and QTF in adults (? ? 0.46) and in chil-
dren (0.44) (Table 2). There was fair agreement between
T-SPOT.TB and QTF in both adults and children (? ?
0.33); per cent agreement ranged from 48% to 79%.
Of the 20 subjects who had discordant results, 13
(65%) were children (Table 3). In 42% of subjects with
discordant results, IGRA values were close to the
manufacturers’ recommended cut-offs. QTF nil values
ranged from 0.09 to 0.39 IU/ml in children and from
0.02 to 0.59 IU/ml in adults. In six of 13 subjects with
discordance between QTF and T-SPOT.TB and/or TST,
the QTF result was negative after subtraction of the
nil response. In all four subjects with a positive TST
and negative T-SPOT.TB, the T-SPOT.TB value was
negative after subtraction of the nil response.
Clinical characteristics and measures of M. tuberculosis infection in HIV-infected adults and children
(N ? 43)
(n ? 23)
(n ? 20)
n/N (%) Mean (SD)
n/N (%) Mean (SD) n/N (%)
Patients with reported TB contact
Patients receiving ART*
18.7 (16.6)4.4 (3.3)35.2 (7.7)
11/43 (25.6)11/23 (48.0)0/20 (0.0)
*Significant difference between adults and children.
†TST ?5 mm was classified as positive.
‡Failed phlebotomy; 10 ml blood was not obtained in children; T-SPOT.TB was preferentially completed in children when ?10 ml blood available.
Significant difference (?2, P ? 0.05) in rates of positive results across tests:
TST and QTF in entire sample
TST and QTF in sub-sample of adults
T-SPOT.TB and QTF in entire sample
T-SPOT.TB and QTF in sub-sample of children
T-SPOT.TB and QTF in sub-sample of adults
SD ? standard deviation; BCG ? bacille Calmette-Guérin; TB ? tuberculosis; ART ? antiretroviral treatment; TST ? tuberculin skin test; QTF ? Quantiferon-Gold.
TST and IGRA responses in HIV-infected adults and children
TST vs. T-SPOT
TST vs. QTF
T-SPOT vs. QTF
* TST positive ?5 mm.
TST ? tuberculin skin test; IGRA ? interferon-? release assays; HIV ? human immunodeficiency virus; T-SPOT ? T-SPOT.TB; QTF ? Quantiferon-Gold.
The International Journal of Tuberculosis and Lung Disease
When controlling for age and CD4 count, logistic
regression found no association between the TST or
IGRAs and M. tuberculosis exposure. In these models,
only age and TST were associated (odds ratio [OR]
1.06, 95% confidence interval [CI] 1.003–1.11). In
exploratory analysis using linear regression and con-
tinuous expression of IGRA outcomes while control-
ling for CD4 count, the QTF ESAT-6 and CFP-10 re-
sponses in adults with known M. tuberculosis exposure
were respectively 2.5 and 1.5 standard deviations higher
than the responses in adults with no known M. tubercu-
losis exposure (P ? 0.0004, 0.04) (Table 4). When sim-
ilarly controlling for CD4 percentage in children, there
was no association between known M. tuberculosis
exposure and QTF responses. Linear regression models
did not find an association between CD4 count or per
cent and the magnitude of any IGRA response.
This is the first study to report on agreement between
both IGRAs and the TST in HIV-infected children.
Discordant results between the TST and IGRA responses
TST, mm T-SPOT QTF
T-SPOT spot count
n (%)Panel A* Panel B* NilESAT-6*CFP-10*Nil
*Values listed indicate the antigen response minus the negative (nil) control response. When the negative control response was greater than the ESAT-6 or CFP-10
response, the resulting value is indicated as negative.
†The nil control had ?6 spots. The value indicated is the difference between the test antigen spot count and twice the negative control spot count, as recom-
mended by the manufacturers.
‡In 3/12 individuals with discordant TST, T-SPOT.TB and QTF results, the QTF results obtained were close to the cut-off values recommended by the manufacturer.
TST ? tuberculin skin test; IGRA ? interferon-gamma release assays; T-SPOT ? T-SPOT.TB; QTF ? Quantiferon-Gold; ESAT-6 ? early secretory antigenic target-6;
CFP-10 ? culture filtrate protein-10.
CD4 count in HIV-infected adults and children
Linear regression for continuous measures of ESAT-6 and CFP-10 responses in relation to M. tuberculosis exposure and
Continuous ESAT-6 and CFP-10 measures of IGRA test responses
Beta† (P value)
Beta† (P value)
Beta† (P value)
Beta† (P value)
Covariates Unadjusted Adjusted‡
Adults (n ? 19)
Children (n ? 12)
CD4 per cent
?0.02 (0.47) ?0.02 (0.52)
?0.05 (0.36) ?0.03 (0.55)
—0.004 (0.87)0.03 (0.40)
*Outcome measures were log transformed and standardized with a mean ? 0 and standard deviation ? 1.
†Beta ? regression coefficient.
‡Linear regression adjusted for CD4 count in adults and CD4 per cent in children.
ESAT-6 ? early secretory antigenic target; CFP-10 ? culture filtrate protein 10 kDa; HIV ? human immunodeficiency virus; IGRA ? interferon-? release assays;
T-SPOT ? T-SPOT.TB; QTF ? Quantiferon-Gold.
IGRA discordance in HIV-infected individuals
Measures of M. tuberculosis infection were more often
positive in adults than in children, demonstrating the
age-dependent risk of acquiring M. tuberculosis in-
fection in endemic settings and the high likelihood of
remote infection in adults. The proportion of individ-
uals in whom M. tuberculosis infection was identified
varied significantly between the TST, T-SPOT.TB and
QTF, with the highest proportion of positive tests
identified by T-SPOT.TB. Discordant results between
all three tests were common. Indeterminate IGRA re-
sults were found in adults who had low CD4 counts.
Among HIV-infected adults attending a routine
HIV out-patient clinic, respectively 62.5%, 72.2% and
35.3% were positive by TST, T-SPOT.TB and QTF. A
recent out-patient based study from a neighbouring
community found 83%, 59% and 46% of non-HIV-
infected adults to be positive by TST, T-SPOT.TB and
QTF.18 We expect that at least a similar proportion of
HIV-infected and non-HIV-infected adults would have
positive test results based on the high risk of M. tuber-
culosis infection in HIV-infected individuals in endemic
settings. Historical comparison of our results in HIV-
infected adults with the results of non-HIV-infected
adults reported in the study discussed above suggests
that T-SPOT.TB may be a sensitive measure of
M. tuberculosis infection in HIV-infected adults.
Although QTF and TST had moderate levels of
agreement, QTF and T-SPOT.TB and TST and
T-SPOT.TB had only fair agreement when compared
in adults and children collectively. Agreement between
tests in adults was similar to agreement demonstrated
in previous studies of HIV-infected adults.18 Agreement
between the TST and T-SPOT.TB was age-dependent,
with agreement being fair in adults and poor in chil-
dren. Age stratification did not affect the agreement
between other test comparisons. The effect of age on
test agreement may be explained by differences in the
tests’ abilities to detect recent infection; for example,
the T-SPOT.TB may be more accurate in detecting re-
cent infection, as is likely to be observed in young chil-
dren. The influence of age on IGRA outcomes may
also reflect age-related differences in immune matura-
tion and IFN-? production.19
The lack of a gold standard for the detection of
M. tuberculosis infection complicates the direct compar-
ison of IGRAs and the TST. We found no sets of discor-
dant results with a positive QTF and negative TST or
T-SPOT.TB, suggesting that the QTF test may lack
sensitivity compared to the TST and T-SPOT.TB in
HIV-infected individuals. In contrast, T-SPOT.TB
identified a significantly higher proportion of infected
individuals than QTF, which suggests that it may be
less specific than QTF in detecting M. tuberculosis
infection in HIV-infected individuals. Results of
T-SPOT.TB in our study indicate an exceedingly high
risk of M. tuberculosis infection in HIV-infected adults
and children and that M. tuberculosis infection in HIV-
infected individuals may be underestimated by QTF.
Observed discrepancies between IGRAs may be re-
lated to several technical and interpretation aspects
related to test methodology. For T-SPOT.TB, the same
numbers of peripheral blood mononuclear cells are
added to wells, thereby correcting for the differences
in T-cell count, whereas for QTF, the identical volume
of blood is added without correction for variability in
T-cell count. In many of the subjects with discordant
results, IGRA results were close to recommended cut-
off values. Although the range of the nil values indi-
cated limited background IFN-? response, relatively
small responses to the nil control appear to impact
test interpretation in subjects with low levels of IFN-?
production. Furthermore, the selection of the IGRA
cut-off point may also affect test interpretation; low
cut-off points may be a possible cause of lower spec-
ificity, higher rates of false positives and poor PPV.
This may be relevant to specific subpopulations, where
different cut-off point values may be required.
In contrast to a previous study where 97% of HIV-
infected subjects had evaluable T-SPOT.TB test re-
sults,10 20% of HIV-infected adults in our study had
one or more indeterminate IGRA results. No indeter-
minate results were observed in children, who had
higher CD4 counts than adults. Consistent with pre-
vious studies of IGRA performance in immune sup-
pressed patients,20,21 adults with indeterminate IGRA
results in our study tended to have low CD4 counts
and negative TST responses. These findings not only
highlight the risk of TST anergy in HIV-infected indi-
viduals with low CD4 counts, but also suggest that
immune suppression may compromise the NPV of
IGRAs in this group.
In contrast to previous household contact studies in
high-burden settings22,23 and contact tracing studies
in low-burden settings,24–26 we found no association
between IGRA responses and M. tuberculosis expo-
sure in our logistic regression. Linear regression dem-
onstrated an association between M. tuberculosis ex-
posure and QTF antigen responses in adults, but no
association between exposure and T-SPOT.TB anti-
gen responses. We suspect that this lack of association
reflects the negative impact of immune suppression on
test performance. As adults with low CD4 counts were
more likely to have had indeterminate QTF results re-
sulting in exclusion from regression analysis, the im-
pact of immune suppression on QTF test performance
was limited. This lack of association may also be re-
lated to the time since exposure in our subjects. Only
20% of our adult subjects had documented recent
M. tuberculosis exposure, but up to 72% had a positive
measure of M. tuberculosis infection. A significant pro-
portion of adults were therefore likely to have had re-
mote infection. These findings highlight the potential
influence of time since infection on the PPVs and NPVs
of IGRAs, even in individuals living in settings highly
endemic for TB.
The main limitations of this study are the small
The International Journal of Tuberculosis and Lung Disease
sample size, its cross-sectional nature and incomplete
QTF testing in children. We were unable to control
for other factors that may have influenced test out-
comes, such as the time since M. tuberculosis infection
Despite the potential of IGRAs to improve the iden-
tification of individuals with M. tuberculosis infec-
tion, there are limited data to guide the interpretation
of IGRAs in HIV-infected adults and children. In chil-
dren and adults with HIV-associated immune suppres-
sion, IGRA test properties may be affected by test cut-
off point and nil control responses. The impact of low
CD4 counts and remote M. tuberculosis infection re-
quire further investigation. Longitudinal assessment
with serial measures is required to elucidate the clini-
cal significance of discordant IGRA and TST results
and the predictive value of IGRAs for M. tuberculosis
infection in HIV-infected individuals.
The authors would like to thank the study nurse, Tygerberg Family
Clinic HIV staff and patients, and especially C Edson and V O’Brien.
This study was supported by the South African National Research
Foundation NRF Thuthuka TTK2006051500016, the Bill and
Melinda Gates Foundation through Grand Challenges in Global
Health Grant 37772 and NIH IK23-HD40982.
1 Dolin P J, Raviglione M C, Kochi A. Global tuberculosis inci-
dence and mortality during 1990–2000. Bull World Health
Organ 1994; 72: 213–220.
2 Nelson L J, Wells C D. Global epidemiology of childhood
tuberculosis. Int J Tuberc Lung Dis 2004; 8: 636–647.
3 Corbett E L, Charalambous S, Moloi V M, et al. Human
immunodeficiency virus and the prevalence of undiagnosed tuber-
culosis in African gold miners. Am J Respir Crit Care Med 2004;
4 Lawn S D, Badri M, Wood R. Tuberculosis among HIV-
infected patients receiving HAART: long term incidence and risk
factors in a South African cohort. AIDS 2005; 19: 2109–2116.
5 Mukadi Y D, Wiktor S Z, Coulibaly I M, et al. Impact of HIV
infection on the development, clinical presentation, and out-
come of tuberculosis among children in Abidjan, Cote d’Ivoire.
AIDS 1997; 11: 1151–1158.
6 Zar H J, Cotton M F, Strauss S, et al. Effect of isoniazid pro-
phylaxis on mortality and incidence of tuberculosis in children
with HIV: randomised controlled trial. BMJ 2007; 334: 136.
7 Menzies D, Pai M, Comstock G. Meta-analysis: new tests for
the diagnosis of latent tuberculosis infection: areas of uncer-
tainty and recommendations for research. Ann Intern Med 2007;
8 Jachuck S J, Bound C L, Price P. The importance of tuberculin
retesting in the adult community at risk. J R Coll Gen Pract
1984; 34: 210–212.
9 Liebeschuetz S, Bamber S, Ewer K, Deeks J, Pathan A A,
Lalvani A. Diagnosis of tuberculosis in South African children
with a T-cell-based assay: a prospective cohort study. Lancet
2004; 364: 2196–2203.
10 Dheda K, Lalvani A, Miller R F, et al. Performance of a T-cell-
based diagnostic test for tuberculosis infection in HIV-infected
individuals is independent of CD4 cell count. AIDS 2005; 19:
11 Chapman A L, Munkanta M, Wilkinson K A, et al. Rapid
detection of active and latent tuberculosis infection in HIV-
positive individuals by enumeration of Mycobacterium tuber-
culosis-specific T cells. AIDS 2002; 16: 2285–2293.
12 den Boon S, Verver S, Marais B J, et al. Association between
passive smoking and infection with Mycobacterium tuberculosis
in children. Pediatrics 2007; 119: 734–739.
13 Marais B J, Hesseling A C, Gie R P, Schaaf H S, Beyers N. The
burden of childhood tuberculosis and the accuracy of community-
based surveillance data. Int J Tuberc Lung Dis 2006; 10: 259–
14 South Africa Department of Health. National HIV and syphilis
antenatal seroprevalence survey in South Africa 2005. Pretoria,
South Africa: DoH, 2006.
15 Jordan T J, Sunderam G, Thomas L, Reichman L. Tuberculin
reaction size measurements by the pen method compared to
traditional palpation. Chest 1987; 92: 234–236.
16 Bouros D, Zeros G, Panaretos C, Vassilatos C, Siafakas N. Pal-
pation vs. pen method for the measurement of skin tuberculin
reaction (Mantoux test). Chest 1989; 99: 416–419.
17 Pai M, Lewinsohn D M. Interferon-gamma assays for tubercu-
losis: is anergy the Achilles’ heel? Am J Respir Crit Care Med
2005; 172: 519–521.
18 Rangaka M X, Wilkinson K A, Seldon R, et al. Effect of HIV-1
infection on t-cell-based and skin test detection of tuberculosis
infection. Am J Respir Crit Care Med 2007; 175: 514–520.
19 Kampmann B, Tena-Coki G, Anderson S. Blood tests for diagno-
sis of tuberculosis. Lancet 2006; 368: 282, author reply 282–283.
20 Ferrara G, Losi M, D’Amico R, et al. Use in routine clinical
practice of two commercial blood tests for diagnosis of infec-
tion with Mycobacterium tuberculosis: a prospective study.
Lancet 2006; 367: 1328–1334.
21 Ferrara G, Losi M, Meacci M, et al. Routine hospital use of a
new commercial whole blood interferon-gamma assay for the
diagnosis of tuberculosis infection. Am J Respir Crit Care Med
2005; 172: 631–635.
22 Hill P C, Brookes R H, Adetifa I M, et al. Comparison of
enzyme-linked immunospot assay and tuberculin skin test in
healthy children exposed to Mycobacterium tuberculosis. Pedi-
atrics 2006; 117: 1542–1548.
23 Hill P C, Brookes R H, Fox A, et al. Large-scale evaluation of
enzyme-linked immunospot assay and skin test for diagnosis of
Mycobacterium tuberculosis infection against a gradient of ex-
posure in The Gambia. Clin Infect Dis 2004; 38: 966–973.
24 Shams H, Weis S E, Klucar P, et al. Enzyme-linked immunospot
and tuberculin skin testing to detect latent tuberculosis infec-
tion. Am J Respir Crit Care Med 2005; 172: 1161–1168.
25 Ewer K, Deeks J, Alvarez L, et al. Comparison of T-cell-based
assay with tuberculin skin test for diagnosis of Mycobacterium
tuberculosis infection in a school tuberculosis outbreak.
Lancet 2003; 361: 1168–1173.
26 Lalvani A, Pathan A A, Durkan H, et al. Enhanced contact
tracing and spatial tracking of Mycobacterium tuberculosis in-
fection by enumeration of antigen-specific T cells. Lancet 2001;
IGRA discordance in HIV-infected individuals
R É S U M É
CONTEXTE : District de Tygerberg, Province Western
Cape, Afrique du Sud.
OBJECTIF : Mesurer la concordance de deux tests de
libération d’interferon-gamma (IGRA) avec le test cutané
tuberculinique (TST) pour la détection de l’infection par
Mycobacterium tuberculosis chez les adultes et les enfants
infectés par le virus de l’immunodéficience humaine (VIH)
dans un contexte de haute incidence de tuberculose (TB).
SCHÉMA : Etude transversale.
RÉSULTATS : Chez 20 adultes et 23 enfants infectés par
le VIH, les tests ont donné des résultats discordants
comportant des tests positifs pour 61% des cas avec
T-SPOT.TB, 41% avec TST et 28% avec Quanti-
FERON® Gold (QTF). Chez les enfants, la concordance
entre le TST et le T-SPOT.TB est médiocre (? ? 0,02),
mais elle est modérée entre le TST et le QTF (? ? 0,44).
Chez les adultes, il y a une concordance modérée entre le
TST et le T-SPOT.TB (? ? 0,43) et entre le TST et le QTF
(? ? 0,46). Il y a une bonne concordance entre le
T-SPOT.TB et le QTF tant chez les enfants que chez les
adultes (? ? 0,33). Un ou plusieurs résultats indéter-
minés de l’IGRA se rencontrent chez 20% des adultes.
CONCLUSIONS : Chez les adultes et les enfants infectés
par le VIH, la concordance entre le TST et les IGRA est
de médiocre à modérée. Le T-SPOT.TB pourrait avoir
une meilleure sensibilité pour la détection de l’infection
par M. tuberculosis chez les individus infectés par le VIH
par comparaison avec le QTF et le TST. Chez les indivi-
dus infectés par le VIH, les propriétés du test IGRA sont
influencées par la limite de positivité du test et les
réponses négatives des contrôles.
R E S U M E N
MARCO DE REFERENCIA : Distrito de Tygerberg, en la
provincia del Cabo Occidental en Sudáfrica.
OBJETIVO : Evaluar la concordancia entre dos pruebas
de liberación de interferón-gamma (IGRA) y la reacción
cutánea a la tuberculina (TST) en la detección de la infec-
ción por Mycobacterium tuberculosis en niños y adultos
infectados por el virus de la inmunodeficiencia humana
(VIH), en un medio con alta endemia de tuberculosis (TB).
MÉTODOS : Estudio transversal.
RESULTADOS : En pacientes adultos (n ? 20) y niños
(n ? 23) infectados por el VIH, las pruebas dieron re-
sultados discordantes con 61% de resultados positivos
con T-SPOT.TB, 41% con la TST y 28% con el Quanti-
FERON® TB Gold (QTF). En los niños se observó una
concordancia baja entre la TST y la prueba T-SPOT.TB
(? ? ?0,02) y fue moderada entre la TST y el QTF (? ?
0,44). En los adultos se encontró concordancia moderada
entre la TST y la T-SPOT.TB (? ? 0,43) y entre la TST
y el QTF (? ? 0,46). La concordancia entre T-SPOT.TB
y QTF fue aceptable (? ? 0,33). Veinte por ciento de
los adultos presentó uno o varios resultados indetermi-
nados con las pruebas de IGRA.
CONCLUSIÓNES : Existe concordancia de baja a moderada
entre la TST y las pruebas de IGRA en niños y adultos
infectados por el VIH. La prueba T-SPOT.TB podría ex-
hibir una mayor sensibilidad en la detección de la infec-
ción tuberculosa en personas infectadas por el VIH, en
comparación con el QTF y la TST. En estas personas,
las propiedades de las pruebas de IGRA se ven afectadas
por el umbral discriminatorio de la prueba y la respuesta
a los controles negativos.