Accepted Manuscript

Detection of Interleukin-2 is not useful for distinguishing between latent and active
tuberculosis in clinical practice: A prospective cohort study

Miguel Santin, Francisco Morandeira-Rego, Fernando Alcaide, Ramón Rabuñal,

Luís Anibarro, Ramón Agüero-Balbín, Xavier Casas-Garcia, Elvira Pérez-Escolano,
Maria D. Navarro, Francisca Sánchez, Amparo Coira-Nieto, Matilde Trigo-Daporta,
Amaya Martinez-Meñaca, Araceli Gonzalez-Cuevas, Maria D. López-Prieto, Ángel
Domínguez-Castellano, Neus Jové

PII: S1198-743X(16)30390-1
DOI: 10.1016/j.cmi.2016.09.004
Reference: CMI 716

To appear in: Clinical Microbiology and Infection

Received Date: 18 July 2016

Revised Date: 10 September 2016
Accepted Date: 11 September 2016

Please cite this article as: Santin M, Morandeira-Rego F, Alcaide F, Rabuñal R, Anibarro L, Agüero-
Balbín R, Casas-Garcia X, Pérez-Escolano E, Navarro MD, Sánchez F, Coira-Nieto A, Trigo-Daporta
M, Martinez-Meñaca A, Gonzalez-Cuevas A, López-Prieto MD, Domínguez-Castellano Á, Jové N,
Detection of Interleukin-2 is not useful for distinguishing between latent and active tuberculosis in
clinical practice: A prospective cohort study, Clinical Microbiology and Infection (2016), doi: 10.1016/
j.cmi.2016.09.004.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Category: Research Note

Title:

Detection of Interleukin-2 is not useful for distinguishing between latent and

active tuberculosis in clinical practice: A prospective cohort study

Authors:

PT
Miguel Santin

RI
Service of Infectious Diseases, Bellvitge University Hospital-IDIBELL,

L’Hospitalet de Llobregat, Barcelona, Spain; Department of Clinical Sciences,

SC
University of Barcelona, L’Hospitalet de Llobregat, Barcelona, Spain; SEIMC

Study Group for Mycobacterial Infections (GEIM).

U
AN
Francisco Morandeira-Rego

Service of Immunology, Bellvitge University Hospital-IDIBELL, L’Hospitalet de

M

Llobregat, Barcelona, Spain.

Fernando Alcaide
D

Service of Microbiology, Bellvitge University Hospital-IDIBELL, L’Hospitalet de

TE

Llobregat, Barcelona, Spain; Department of Pathology and Experimental

EP

Therapy, University of Barcelona, L’Hospitalet de Llobregat, Barcelona, Spain;

SEIMC Study Group for Mycobacterial Infections (GEIM).

C

Ramón Rabuñal
AC

Unit of Infectious Diseases, Hospital Universitario Lucus Augusti, Lugo, Spain

Luís Anibarro

Unit of Tuberculosis, Service of Infectious Diseases-Internal Medicine,

Complexo Hospitalario Universitario de Pontevedra, Pontevedra, Spain

Ramón Agüero-Balbín

1
 ACCEPTED MANUSCRIPT
Service of Respiratory Diseases, Hospital Universitario Marqués de Valdecilla,

Santander, Spain

Xavier Casas-Garcia

Respiratory Diseases, Hospital General Parc Sanitari Sant Joan de Déu, Sant

Boi de Llobregat, Barcelona

PT
Elvira Pérez-Escolano

RI
Unidad de Gestión Clínica Enfermedades Infecciosas y Microbiología, Hospital

de Jerez, Jerez de la Frontera, Cádiz, Spain

SC
Maria D. Navarro

Clinic Unit of Infectious Diseases and Preventive Medicine, Hospital Virgen

Macarena, Sevilla, Spain

U
AN
Francisca Sánchez
M

Service of Infectious Diseases, Institut Mar d’Investigacions Mèdiques (IMIM),

Hospital del Mar, Barcelona, Spain

D

Amparo Coira-Nieto
TE

Service of Microbiology, Hospital Universitario Lucus Augusti, Lugo, Spain

Matilde Trigo-Daporta
EP

Service of Microbiology, Complexo Hospitalario Universitario de Pontevedra,

Pontevedra, Spain
C

Amaya Martinez-Meñaca
AC

Service of Respiratory Diseases, Hospital Universitario Marqués de Valdecilla,

Santander, Spain

Araceli Gonzalez-Cuevas

Service of Microbiology, Hospital General Parc Sanitari Sant Joan de Déu, Sant

Boi, Barcelona, Spain

2
 ACCEPTED MANUSCRIPT
Maria D. López-Prieto

Unidad de Gestión Clínica Enfermedades Infecciosas y Microbiología, Hospital

de Jerez, Jerez de la Frontera, Cádiz, Spain

Ángel Domínguez-Castellano

Clinic Unit of Infectious Diseases and Preventive Medicine, Hospital Virgen

PT
Macarena, Sevilla, Spain

RI
Neus Jové

Service of Infectious Diseases, Institut Mar d’Investigacions Mèdiques (IMIM),

SC
Hospital del Mar, Barcelona, Spain

Running title: IL-2 to distinguish latent from active tuberculosis

U
AN
Word count: Abstract (250), paper (1280)

Corresponding author:
M

Miguel Santin

Service of Infectious Diseases, Bellvitge University Hospital-IDIBELL

D

08907 L’Hospitalet de Llobregat, Barcelona, Spain

TE

Phone: +34932607625
EP

Fax: +34932607637

Mail: msantin@bellvitgehospital.cat
C
AC

3
 ACCEPTED MANUSCRIPT

Abstract

Objective: Previous reports have identified interleukin-2 (IL-2), quantified in the

supernatants of QuantiFERON®-TB Gold In-tube (QFT) after 72 hours of

incubation, as a potential biomarker for distinguishing between latent and active

PT
tuberculosis (TB). However, its validity has not been tested in an appropriate

RI
clinical cohort.

Methods: A multicentre study of 161 consecutive adult patients undergoing

SC
evaluation for active TB at eight TB Units in Spain. IFN-γ and IL-2 were

assessed in the supernatant of QFT after 16-24 h and 72 h incubation. The

U
accuracy of IL-2 for indicating latent TB infection (LTBI) was assessed by
AN
receiving operating characteristic (ROC) curves. .
M

Results: Twenty-eight participants were not infected, 43 had LTBI, 69 had TB,

and 21 were not classifiable. Median (interquartile range [IQR]) interleukin-2

D

concentrations after 72 h of incubation were 0.0 pg/ml (0.0-0.0) in uninfected

TE

individuals, 261.0 pg/ml (81.0-853.0) in LTBI individuals, 166.5 pg/ml (33.5-

551.5) in extra-pulmonary TB patients, 95.0 pg/ml (26.0-283.0) in smear-

EP

negative pulmonary TB patients, and 38.5 pg/ml (7.5-178.0) in smear-positive

pulmonary TB patients (p<0.0001). The area under the curve (AUC) of the ROC
C

curve (95% confidence interval [95%CI]) of IL-2 after 72 h of incubation for the
AC

diagnosis of LTBI was 0.63 (0.53-0.74) when all TB cases were considered as a

single group, ranging from 0.59 (0.47-0.71) to 0.72 (0.58-0.85) when only extra-

pulmonary and smear-positive pulmonary TB cases respectively were

considered.

4
 ACCEPTED MANUSCRIPT
Conclusions: Quantification of IL-2 in the supernatant of QTF after a prolonged

incubation is not useful to distinguish between LTBI and active disease in

clinical practice.

Keywords: Interleukin-2, IFN-γ, latent tuberculosis infection, active

tuberculosis.

PT
RI
U SC
AN
M
D
TE
C EP
AC

5
 ACCEPTED MANUSCRIPT
Introduction

Although interferon-gamma (IFN-γ) release assays (IGRA) have improved the

diagnosis of tuberculosis (TB) infection, they are unable to distinguish between

latent (LTBI) and active TB [1]. In recent years, many efforts have been made to

determine the usefulness of immunologic biomarkers of LTBI and active TB

PT
diagnosis [2]. The rationale is the fact that Mycobacterium tuberculosis specific

RI
T-cell response and its cytokine expression vary depending on the clinical state

of the infection [3,4]. Interleukin-2 (IL-2) is one of the most extensively

SC
evaluated, because IL-2- and IL-2/IFN-γ-secreting central memory T (TCM) cells

predominate in LTBI, while IFN-γ-secreting effector memory T (TEM) cells prevail

U
in active TB (5). Moreover TEM cells proliferate within 24 h whereas TCM need a
AN
few days [5]. In this regard, two previous studies reported higher levels of IL-2 in
M

latently infected individuals compared with TB patients in M. tuberculosis-

specific-stimulated supernatants after prolonged incubation (72 h) of

D

QuantiFERON®-TB Gold In-tube (QFT), suggesting that quantification of IL-2 at

TE

72 h of incubation of QFT may distinguish between LTBI and active TB disease

[6,7]. However, these preliminary promising results have not been validated yet.
EP

In this study we aimed to validate the quantification of IL-2 in the

supernatant of QFT after 72 h incubation in a cohort of patients with suspected

C

TB for use in clinical practice to distinguish between LTBI and active TB.
AC

Methods

We conducted a prospective, multicentre study of consecutive adult patients

undergoing evaluation for active TB at eight TB Units in Spain (October 2013 to

December 2015). Patients ≥18 years with suspicion of active pulmonary or

extra-pulmonary TB, for whom the treating clinician considered ordering an

6
 ACCEPTED MANUSCRIPT
immunodiagnostic test for TB infection as a part of the diagnostic workup, were

eligible for the study. Exclusion criteria were immunosuppression, known

previous latent or active TB, and a positive nucleic amplification test (NAT) from

any clinical sample.. For each patient, two sets of QuantiFERON®-TB Gold In-

tube (Cellestis/Qiagen, Carnegie, Australia/Germany) were processed,

PT
incubated for 16-24 h (24h-QFT) according to the manufacturer’s instructions

RI
and for 72 h (72h-QFT) respectively. The remaining supernatants were

harvested at -70 degrees Celsius for batched analysis of IL-2 at the coordinator

SC
site (BUH). Levels of IL-2 were measured with the commercial ELISA assay

Quantikine®ELISA Human IL-2 Immunoassay (R&D Systems Inc, Minneapolis,

U
MN, USA) according to the manufacturer’s instructions. The investigators who
AN
performed the IFN-γ (QFT) and IL-2 were blinded to the clinical data and
M

tuberculin skin test (TST) and 72h-QFT results of participants.

Diagnosis of TB was considered definite if it was microbiology-confirmed,

D

and probable if, in the absence of microbiological confirmation, compatible

TE

clinical, radiological and/or other supporting laboratory data (adenosine

desaminase [ADA] and/or histology with necrotic granulomatous inflammation)

EP

were present and cure was achieved within six months of specific therapy. LTBI
C

was defined as definite if 24h-QFT and TST (≥5 mm) were positive, and
AC

probable if QFT was positive and TST negative, in absence of active TB.

Patients negative for both TST and QFT were classified as non-infected.

Patients with discordant results (TST+/QFT-) were classified as discordant and

were not included for analyses in any of the other groups. Clinicians who

established the diagnosis were blind to the results of the 72h-QFT.

7
 ACCEPTED MANUSCRIPT
The central point of the analysis was the sensitivity and specificity of the

levels of IL-2 for diagnosing LTBI. We calculated that a sample of 295

individuals would be needed to demonstrate a sensitivity and specificity of 0.90

and 0.95 respectively, assuming a prevalence of LTBI of 50% and 5% of losses

[8]. The funding organism (Instituto de Salud Carlos III, Spanish Ministry of

PT
Health) asked to perform an interim analysis in the middle of the inclusion

RI
period in order to decide whether to extend the period and funding of the study.

Recruitment was discontinued after the analysis of the first 161 valid cases

SC
revealed poor discriminatory power of IL-2 for latent and active TB. After

assessing the distribution of continuous variables with the Kolmogorov-Smirnov

test, they
U
were compared using the Kruskal-Wallis test for independent
AN
samples and the Wilcoxon test for related samples. Categorical variables were
M

compared with the chi-square test. Receiving operating characteristic (ROC)

analyses were performed, plotting sensitivity against specificity at all value

D

thresholds in the data set. Analyses were performed with SPSS statistical
TE

software (version 15.0, SPSS Institute Inc, Chicago, Illinois, USA) and Prism V5

(GraphPad Software Inc, La Jolla, CA, USA). The study was designed and
EP

developed following the QUADAS (Quality Assessment of Diagnostic Accuracy

Studies) criteria [9].

C

Written informed consent was obtained from all participants, and the
AC

ethical committees at the eight sites approved the study.

Results

Two hundred and twenty-one individuals were recruited and signed informed

consent, but 60 were excluded from the analysis. The flowchart of the study

shows the number and causes of exclusion (Figure 1). . Table 1 shows the

8
 ACCEPTED MANUSCRIPT
concentrations of IFN-γ and IL-2, and variations from 24 to 72 h of incubation.

Median (interquartile range [IQR]) background-corrected concentrations of IFN-

γ concentrations after 24 and 72 h of incubationwere higher in M. tuberculosis-

infected subjects (4.26 [0.77-10.34] and 4.55 [0.93-9.61] for 24 and 72 h

respectively) than in uninfected (0.01 [0.0-0.09] and 0.01 [0.0-0.1] for 24 and 72

PT
h respectively) and discordant individuals (0.0 [0.0-0.11] and 0.0 [0.0-0.14] for

RI
24 and 72 h) (p<0.0001). Likewise, the IL-2 concentrations were higher in M.

tuberculosis-infected subjects (98.7 [20.5-439.5] and 169.5 [33.0-581.0] for 24

SC
and 72 h respectively) than in uninfected (0.0 [0.0-0.0] and 0.0 [0.0-0.0] for 24

and 72 h respectively) and discordant individuals (0.0 [0.0-19.0] and 9.4 [0.0-

U
62.7] for 24 and 72 h respectively) (p<0.0001). While there was a significant
AN
increase in the IL-2 concentrations at 72 h with respect to 24 h in both LTBI and
M

TB groups (52 [1.0-276] and 14 [0.0-79] respectively) (p<0.0001),

concentrations of IFN-γ did not change (0.0 [-0.53-0.2]; p=0.99, for LTBI group;
D

and 0.0 [-0.37-0.68]; p=0.46, for TB group). The concentrations of IL-2 at 72 h

TE

of incubation varied across the spectrum of TB infection: 261 pg/ml (IQR 81.0-

853.0) for LTBI, 166.5 pg/ml (IQR 33.5-551.5) for extra-pulmonary TB, 95.0
EP

pg/ml (IQR 26.0-283.0) for smear-negative pulmonary TB, and 38.5 pg/ml (IQR
C

7.5-178.0) for smear-positive TB (p<0.0001).

AC

The area under the curve (AUC) of the ROC curve of the IL-2 levels at 72 h for

the diagnosis of ITBL was 0.63 (95% Confidence Interval [CI] 0.53-0.7;

p=0.018) (Figure 2), ranging from 0.59 (95% CI 0.47-0.71; p=0.16) to 0.72 (95%

CI 0.58-0.85; p=0.006) when only extra-pulmonary and smear-positive TB

pulmonary cases were considered respectively (Figure 3).

Discussion

9
 ACCEPTED MANUSCRIPT
We were unable to confirm the potential utility of IL-2 as a biomarker to

distinguish latent infection from active disease in patients undergoing evaluation

for active TB disease. In agreement with Biselli et al. [6], we observed a

significant increase in IL-2 concentrations after 72 h of incubation, but LTBI and

TB groups presented a considerable overlap. The resulting diagnostic accuracy

PT
was poor, with an AUC of the ROC curve of 0.63, much lower than previously

RI
reported figures [6,7]. Interestingly, the diagnostic accuracy for LTBI, although

remaining suboptimal, improved slightly as the bacillary load of active TB

SC
increased, from extra-pulmonary TB, the lowest, to smear-positive pulmonary

TB, the highest. From the point of view of clinical practice, however, this finding

U
has limited relevance since a diagnostic test of this kind would only be really
AN
useful in paucibacillary difficult-to-diagnose TB cases.
M

There is no obvious explanation for the notable discordance between the

results of these two previous studies and ours. Since our results showed better
D

performance of the test when only smear-positive pulmonary TB cases were

TE

considered, and since previous studies only included confirmed-pulmonary TB

cases, this difference in design could explain, at least in part, the differences
EP

between our results and those of previous studies. In any case, validating the

performance of diagnostic tests in appropriate cohorts is mandatory. Our study

C

was designed and conducted in accordance with current guidelines [9].

AC

Specifically, the study population comprised patients with suspicion of TB,

clinical investigators were blinded to the 72h-QFT and IL-2 results, and

determination of IL-2 was centralised at a single centre to avoid variability.

10
 ACCEPTED MANUSCRIPT
In conclusion, quantification of IL-2 in the supernatant of QFT after a

prolonged incubation is not useful for distinguishing latent from active TB

disease in clinical practice.

Acknowledgements

Collaborators of the IL-2 Study Team

PT
Edu A. Struzka

RI
Service of Microbiology, Bellvitge University Hospital-IDIBELL, L’Hospitalet de

Llobregat, Barcelona, Spain

SC
Joan Climent

Service of Immunology, Bellvitge University Hospital-IDIBELL, L’Hospitalet de

U
AN
Llobregat, Barcelona, Spain.

Antón Penas-Truque.
M

Unidad de Tuberculosis. Hospital Universitario Lucus Augusti, Lugo

Abel Pallarés-Sanmartín
D

Service of Respiratory Diseases, Complexo Hospitalario Universitario de

TE

Pontevedra, Pontevedra, Spain

Mónica Ríos
EP

Service of Internal Medicine, Complexo Hospitalario Universitario de

C

Pontevedra, Pontevedra, Spain

AC

Diego J. Pargada-Ferrer

Service of Respiratory Diseases, Hospital de Barbastro, Huesca, Spain

Jesús Agüero-Balbin

Service of Microbiology, Hospital Universitario Marqués de Valdecilla,

Santander, Spain

Juan F. Rodríguez-Gutiérrez

11
 ACCEPTED MANUSCRIPT
Service of Immunology, Hospital de Jerez, Jerez de la Frontera, Cádiz, Spain

References

1. Santin M, García-García J-M, Domínguez J. Guidelines for the use of

interferon-γ release assays in the diagnosis of tuberculosis infection.

Enferm Infecc Microbiol Clin 2016;34:1–13.

PT
2. Chegou NN, Heyckendorf J, Walzl G, Lange C, Ruhwald M. Beyond the

RI
IFN-γ horizon: Biomarkers for immunodiagnosis of infection with

Mycobacterium tuberculosis. Eur Respir J 2014;43:1472-86.

SC
3. Sallusto F, Geginat J, Lanzavecchia A. Central Memory and Effector

U
Memory T Cell Subsets: Function, Generation, and Maintenance. Annu Rev
AN
Immunol 2004;22:745-63.

4. Lanzavecchia A, Sallusto F. Dynamics of T Lymphocyte Responses:

M

Intermediates, Effectors, and Memory Cells. Science 2000;290:92-97.

5. Sargentini V, Mariotti S, Carrara S, Gagliardi MC, Teloni R, Goletti D, et al.

D

Cytometric detection of antigen-specific IFN-gamma/IL-2 secreting cells in

TE

the diagnosis of tuberculosis. BMC Infect Dis 2009;9:99.

6. Biselli R, Mariotti S, Sargentini V, Sauzullo I, Lastilla M, Mengoni F, et al.

EP

Detection of interleukin-2 in addition to interferon-gamma discriminates

C

active tuberculosis patients, latently infected individuals, and controls. Clin

AC

Microbiol Infect 2010;16:1282-4.

7. Mamishi S, Pourakbari B, Marjani M, Bahador A, Mahmoudi S.

Discriminating between latent and active tuberculosis: The role of

interleukin-2 as biomarker. J Infect 2015;70:429-31.

8. Carley, S; Dosman, S; Jones SR; Harrison M. Simple nomograms to

calculate sample size in diagnostic studies. Emerg Med J 2005;22:180-1.

12
 ACCEPTED MANUSCRIPT
9. Oliveira MRF de, Gomes A de C, Toscano CM. QUADAS and STARD:

evaluating the quality of diagnostic accuracy studies. Rev Saude Publica

2011;45:416-22.

PT
Transparency declaration

- M.S. is the PI of a Clinical Trial assessing QuantiFERON®-TB Gold In-tube in

RI
contact-tracing, which was supplied with blood collection tubes by Cellestis, Inc.

SC
(Carnegie, Australia). The other authors declare no conflict of interest.

-The study was funded by the Ministerio de Economía y Competitividad,

U
Instituto de Salud Carlos III (ISCIII) of the Spanish Government (Grant:
AN
PI12/02322).
M

-The study was partially presented at the XX Congress of the Spanish Society

of Infectious Diseases and Clinical Microbiology (SEIMC), Barcelona (Spain),

D

on Saturday, May 28th, 2016.

TE
C EP
AC

13
 ACCEPTED MANUSCRIPT

Table 1. Main characteristics and QTF and IL-2 results of 161 participants

PT
Non-TB-infected Group Discordant Group LTBI Group Active TB Group

n= 28 n= 21 n= 43 n= 69

RI
Age, median, years 57 (44.5-77.3) 49 (44.5-54) 54 (46-64) 41 (31-52)

SC
Male gender, no.(%) 16 (57.1) 10 (47.6) 30 (69.8) 37 (53.6)

U
Foreign born, no.(%) 2 (7.1) 4 (19) 4 (9.3) 21 (30.4)

AN
BCG vaccination, no. (%) 9/27 (33.3) 18/21 (85.7) 21 (100) 25/68 (36.8)

M
Pulmonary TB, no.(%) -- -- -- 37 (53.6)

D
Definite diagnosis, no. (%) -- -- -- 33/37 (89)

Extra-pulmonary TB, no. (%) --

TE -- -- 32 (46.4)
EP
Definite diagnosis, no.(%) -- -- -- 11/32 (34.4)
C

TST ≥5 mm, no.(%) -- 21 (100) 33 (84.6) 48/54 (88.9)

AC

QTF (24 h) positive, no.(%) -- -- 43 (100) 57 (82.6)

IFN-γ (Ag-Nil), median (IQR), U/ml*

 ACCEPTED MANUSCRIPT

-24 h. incubation 0.01 (0.00-0.09) 0.00 (0.00-0.11) 7.74 (1.85-12.69) 2.77 (0.65-9.12)

-72 h. incubation 0.01 (0.00-0.10) 0.00 (0.00-0.14) 6.56 (2.06-12.93) 3.70 (0.59-8.53)

PT
-∆ 24-72 h. incubation 0.00 (0.00-0.06) 0.00 (-0.03-0.03) 0.00 (-0.53-0.23) 0.00 (-0.37-0.68)

RI
IL-2 (Ag-Nil), median (IQR), pg/ml**

SC
-24 h. incubation 0.0 (0.0-0.0) 0.0 (0.0-19.0) 199.0 (62.0-466.0) 63.0 (17.0-331.0)

U
-72 h. incubation 0.0 (0.0-0.0) 9.4 (0.0-62.7) 261.0 (81.0-853.0) 115.0 (20.0-364.0)

AN
-∆ 24-72 h. incubation 0.0 (0.0-0.0) 2.3 (0.0-14.0) 52.0 (1.0-276.5) 14.0 (0.0-79.0)

M
Time until incubation, median (IQR), 1.23 (0.76-1.85) 1.00 (0.42-1.45) 1.25 (0.75-1.88) 0.75 (0.29-1.46)

D
hours (n= 141)

Time of incubation, median (IQR),

TE
EP
hours (n= 141)
C

-24h-QTF 23.73 (22.04-24.00) 24.00 (22.75-24.75) 23.17 (22.54-24.00) 24.00 (22.44-24.00)

AC

-72h-QTF 71.62 (69.48-72.13) 72.00 (71.00-72.75) 71.63 (69.50-62.42) 71.50 (69.54-72.00)

n/N= Number with the condition/number for which the data is available
 ACCEPTED MANUSCRIPT

BCG= Bacillus Calmette-Guérin; QTF= QuantiFERON-TB Gold In-tube; IL-2= interleukin-2; IFN-γ= interferon-γ; TB= tuberculosis; LTBI= latent
tuberculosis infection; Ag= antigen

PT
*Differences in IFN-γ concentrations between the four groups, both at 24 and 72 h of incubation (p<0.0001). Differences between LTBI and TB
at 24 h (p= 0.99) and 72 h (p= 0.46).

RI
** Differences in IL-2 concentrations between the four groups, both at 24 and 72 h of incubation (p<0.0001). Differences between LTBI and TB
at 24 and 72 h (p<0.0001).

SC
IFNAg-Nil; ***Data available for 141 participants

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Recruited
(informed consent signed)
(n= 222)

PT
Excluded

RI
(no inclusion criteria fulfilled)
(n= 11)

SC
Eligible
(n= 211)

U
AN
Excluded
- Indeterminate QFT result (n= 5)

M
- Sample for IL-2 unavailable (n= 20)
- IL-2 not determined (technical problems) (n= 25)

D
Included TE
(n= 161)
C EP
AC

Non-TB Discordant Latent TB Active

infection (TST +/ QFT -) infection TB
(n= 28) (n= 21) (n= 43) (n= 69)

QFT= QuantiFERON®-TB Gold In-tube ; IL-2= Interleukine 2; TB= Tuberculosis

 ACCEPTED MANUSCRIPT

a) IFN- (24 h) IFN- (72 h) ROC curve (IFN- 72 h)

100
60 P<0.0001 60 P<0.0001

80

PT
40 40

Sensitivity (%)
60
U/ml

RI
20 20 40

SC
20

U
0 0
AUC= 0.64 (95%CI 0.53-0.74)

AN
0
1 2 3 4 1 2 3 4 0 20 40 60 80 100
100-Specificity (%)

M
b) IL-2 (24 h) IL-2 (72 h) ROC curve (IL-2 72 h)

D
100
2500 2500
P<0.0001 P<0.0001

2000
TE 80
2000
EP

Sensitivity (%)
pg/ml

1500 1500 60
C
AC

1000 1000 40

500 500 20

AUC= 0.63 (95%CI 0.53-0.74)

0 0 0
1 2 3 4 1 2 3 4 0 20 40 60 80 100
100-Specificity (%)
 ACCEPTED MANUSCRIPT

a) Extra-pulmonary TB b) Pulmonary TB
100 100

80 80

PT
Sensitivity (%)

Sensitivity (%)
60 60

RI
40 40

SC
20 20

U
AUC= 0.59 (95%CI 0.47-0.71) AUC= 0.66 (95%CI 0.54-0.79)

AN
0 0
0 20 40 60 80 100 0 20 40 60 80 100
100-Specificity (%) 100-Specificity (%)

M
c)

D
Smear () pulmonary TB d) Smear () pulmonary TB
100 100
TE
80 80
EP
Sensitivity (%)

Sensitivity (%)
C

60 60
AC

40 40

20 20

AUC= 0.67 (95%CI 0.53-0.81) AUC= 0.72 (95%CI 0.58-0.85)

0 0
0 20 40 60 80 100 0 20 40 60 80 100
100-Specificity (%) 100-Specificity (%)