ORIGINAL ARTICLE

Acta Medica Indonesiana - The Indonesian Journal of Internal Medicine

Use of Pleural Fluid Interferon-gamma Enzyme-linked
Immunospot Assay in the Diagnosis of Pleural Tuberculosis

Tika Adilistya1, Dalima A.W. Astrawinata1, Ujainah Z. Nasir2

1 Department of Clinical Pathology, Cipto Mangunkusumo Hospital, Jakarta, Indonesia.
2 Department of Internal Medicine, Cipto Mangunkusumo Hospital, Jakarta, Indonesia.

Corresponding author:
Tika Adilistya, MD. Department of Clinical Pathology, Cipto Mangunkusumo Hospital. Jl. Diponegoro 71, Jakarta
10430, Indonesia. email: dr.adilistya@gmail.com.

ABSTRAK
Tujuan: untuk melakukan penilaian terhadap uji diagnostik pemeriksaan interferon-gamma release assay

(IGRA) metode enzyme-linked immunospot (ELISPOT), yaitu T-SPOT.TB, untuk deteksi TB pleura menggunakan
spesimen sel mononuklear (MN) cairan pleura. Metode: sebanyak 48 pasien efusi pleura terduga TB dengan
karakteristik cairan pleura eksudatif berdasarkan kriteria Light dan dominasi sel MN lebih dari 50% dilakukan
pemeriksaan T-SPOT.TB, biakan TB media cair Mycobacterial Growth Indicator Tube (MGIT), dan aktivitas
adenosine deaminase (ADA) cairan pleura. Penyebab lain efusi pleura seperti gagal jantung, gagal ginjal,
sirosis hati, dan keganasan telah disingkirkan. Hasil: sebanyak 39 dari 48 subjek (81,25%) menderita TB
pleura berdasarkan biakan MGIT positif dan/atau ADA lebih dari 40 U/L. Dari jumlah tersebut seluruhnya
positif untuk T-SPOT.TB. Hasil uji diagnostik IGRA ELISPOT untuk diagnosis TB pleura adalah sensitivitas
100%, spesifisitas 88,89%, nilai prediksi positif 97,5%, dan nilai prediksi negatif 100%. Kesimpulan: IGRA
ELISPOT menggunakan specimen cairan pleura merupakan metode diagnostik yang cepat dan reliabel sehingga
bermanfaat untuk diagnosis TB pleura khususnya pada daerah endemis TB.

Kata kunci: interferon-γ release assay, ELISPOT, tuberkulosis pleura.

ABSTRACT
Aim: to evaluate the diagnostic value of an interferon-gamma release assay (IGRA) with enzyme-linked

immunospot (ELISPOT) method, T-SPOT.TB, in the diagnosis of pleural TB using pleural fluid mononuclear
cells (PFMC). Methods: forty-eight subjects, presumed to have pleural TB with exudative pleural effusion
by Light’s criteria, dominated by mononuclear cells, had their pleural fluid specimen tested with T-SPOT.
TB, Mycobacterial Growth Indicator Tube (MGIT) culture, and adenosine deaminase (ADA) activity. Other
causes of pleural effusion such as heart failure, renal failure, hepatic cirrhosis, and malignancy were excluded.
Results: the sensitivity, specificity, positive predictive value, and negative predictive value of the IGRA ELISPOT
assay using PFMC for the diagnosis of pleural TB were 100%, 88.89%, 97.5%, and 100%, respectively.
Conclusion: IGRA with ELISPOT method performed on PFMC is useful for a rapid and reliable diagnosis of
pleural TB in clinical practice, especially in area with high TB burden.

Keywords: interferon-γ release assay, ELISPOT, pleural tuberculosis.

Tika Adilistya Acta Med Indones-Indones J Intern Med

INTRODUCTION
Tuberculosis (TB) remains one of the leading

causes of mortality in the world. Although TB
is typically a disease of the lungs, which serves
both as port of entry and also as the major
site of disease manifestation, Mycobacterium
tuberculosis has the ability to disseminate to
various extrapulmonary sites. Tuberculous
pleurisy, or pleural TB, is the second most
common manifestation of extrapulmonary TB and
a common cause of pleural effusion in endemic
TB areas. To date, diagnosis of pleural TB relies
on either insensitive (acid fast bacilli smears),
unspecific (cell count, biochemical levels), or
time consuming (culture) methods often leading
to defer initiation of therapy. The paucity of
bacilli in pleural fluid leads to low sensitivity of
direct bacillary detection such as Ziehl-Neelsen
staining, culture, and also PCR.1-3 Therefore, a
rapid, accurate diagnostic test is urgently needed
for pleural tuberculosis.

In the late 1990s, many studies about
the M. tuberculosis genome have led to the
identification of TB specific antigens. Among
all antigens, early secretory antigenic target
with 6 kDa molecular weight (ESAT-6) and
culture filtrate protein with 10 kDa molecular
weight (CFP10), which appear exclusively in
M. tuberculosis, are the most immunodominant
and also virulency determinant of mycobacteria.
Other mycobacteria, such as M. bovis in
Bacillus Calmette-Guerin vaccine and most
environmental nontuberculous mycobacteria
do not have these specific antigens therefore
the use of ESAT-6 and CFP10 are beneficial for
TB detection because it is more sensitive and
specific.4

During active TB, antigen-specific T
lymphocytes clonally proliferate and are
recruited to the site of active infection. Those
cells will release more interferon-gamma
(IFN-γ) cytokine after rechallenge with TB
specific antigens in vitro. In pleural TB, more
CD4+ T lymphocyte subgroups are found in the
pleural fluid than in peripheral blood, and also
the level of IFN-γ secreted by T lymphocytes in
pleural fluid are more than the level of IFN-γ in
peripheral blood.5

Until now, the commercial platforms of IGRA

are only validated for blood samples. Based on
facts that more antigen-specific T lymphocytes
are found in pleural fluid than those in peripheral
blood, we were interested in evaluating the
clinical utility of an M. tuberculosis antigen-
specific IFN-γ using enzyme-linked immunospot
(ELISPOT) method, performed on pleural fluid
mononuclear cells (PFMC), for the diagnosis of
pleural TB in a setting with high incidence of
TB disease.

METHODS
A diagnostic study was performed in which

48 consecutive patients with presumed TB pleural
effusion were enrolled at Cipto Mangunkusumo
Hospital from May to September 2015. The study
was approved by the Ethical Committee of Cipto
Mangunkusumo Hospital/Faculty of Medicine
Universitas Indonesia.

Subjects
Inclusion criteria were patients aged more

than 18 years old with suspected pleural effusion
based on clinical symptoms and ultrasound
examination, with no history of antituberculosis
medication in the last 6 months. Patients with
heart failure, renal failure, hepatic cirrhosis, and
malignancy were excluded. Written informed
consents were obtained from all the subjects.
Sample size was 48 subjects, determined using
diagnostic test sampling formula.

Pleural fluid TB culture and ADA activity were
used as the gold standard in this diagnostic study.
Based on WHO guidelines of extrapulmonary
TB management, microscopic examination
is the gold standard for diagnostic. But it has
limitations since the sensitivity is very low. So
we combined with pleural fluid ADA activity to
overcome this. In clinical practice in our centre,
as Indonesia is a high TB burden country, when
a clinically suspected pleural TB patients has
elevated pleural fluid ADA activity and other
causes of false elevation have been excluded,
they will be treated with antituberculosis
although TB culture is negative.

One hundred millilitres (75-125 mL) of
pleural fluid was aspirated. First tube (25 mL) was
for routine pleural fluid analysis such as white
blood cell (WBC) count, mononuclear (MN) cell

Vol 48 • Number 1 • January 2016 Use of pleural fluid interferon-gamma enzyme-linked immunospot assay

count, protein and LDH levels). Determination
of exudate was based on Light’s criteria (fluid/
serum protein ratio >0.5, fluid/serum LDH
ratio >0.6, and LDH >2/3 upper limit of normal
serum LDH).6-7 If the results were not exudate,
subjects would be excluded from the study. If
the results were exudate, subsequent steps were
ADA activity measurement, ELISPOT, and
MGIT culture.

White blood cells count was performed using
Sysmex XE-2100. MN cell count was performed
using Wright stained cytocentrifuged slide which
is presently considered as the “gold standard” for
body fluid differential counting.The ELISPOT
assay and MGIT culture was performed within
2 hours after the collection of pleural fluid.

PFMC ELISPOT Assay
The ELISPOT kit (T-SPOT.TB; Oxford

Immunotec Ltd, Oxford, UK) was performed
according to the manufacturer’s instruction with
some modifications in order to have an adequate
and good quality of pleural fluid mononuclear
cells (PFMC). The sample preparation needs
pleural fluid volumes of 50 mL to 100 mL.
Volumes of 100 mL are more likely to provide
sufficient cells although 50 mL are normally
sufficient. Samples should be stored at room
temperature or 2-8°C. They should not be frozen.
Pleural fluid is transferred to a 50-mL Falcon
tube and centrifuged at 465 g for 15 minutes at
room temperature. After centrifugation, check
if there is a good cell pellet at the bottom of the
tube. Pipette the supernatant carefully to ensure
that there is no loss of cells. If the pellet is small,
centrifugation can be repeated. If the pellet is
contaminated with red blood cells, then carry
out a “hypertonic shock” as follows. Resuspend
the pellet in 2 mL sterile or distilled water and
incubate for 50-60 seconds. To restore the
normal osmolarity, add 2x PBS to the suspension.
After that, dilute the suspension with 1x PBS or
RPMI medium to 20-30 mL and centrifuge the
suspension again at 465 g for 15 minutes. If the
pleural fluid is contaminated with many red blood
cells, a density gradient centrifugation similar to
the density gradient centrifugation used in the
standard T-SPOT.TB test can be performed, such
as the Ficoll procedure and Leucosep method. If
the contamination of the pellet with red blood cells

is low, resuspend the pellet in one mL of RPMI
using a pipette and make up to 20-30 m, then
centrifuge again at 465 g for 15 minutes, remove
supernatant carefully. After that, resuspend the
pellet in 0,7 mL AIM-V medium and process the
T-SPOT.TB test according to the manufacturer’s
instruction for peripheral blood mononuclear
cells (PBMC).T-SPOT.TB uses two TB specific
antigens: ESAT-6 and CFP10. Panel A contains
ESAT-6 and Panel B contains CFP10. The
response of stimulated cultures was considered
positive when the test well (Panel A or Panel B)
contained at least six more spots and had twice the
number of spots shown in the control well. The
background number of spots in negative control
wells was below 10 spots per well in all patients.8

TB Culture Using Mycobacterial Growth
Indicator Tube

Twenty five millilitres of pleural fluid was
centrifuged at 3500 g for 15 minutes. The pellet
was treated with an equal volume of NaOH
4% for 15 minutes at room temperature and
neutralized with sterile phosphate buffer. The
suspension was inoculated for isolation of acid
fast bacilli by culture in MGIT (BD Bactec MGIT
960 system).

Determination of ADA Activity in Pleural Fluid
Adenosine deaminase (ADA) activity levels

were detected using an ADA measurement kit
with enzymatic colorimetric method (Mindray
Bio-Medical Electronics, Shenzhen, China) by
following the manufacturer’s instructions.

Statistical Analysis
Data results were analyzed by using MS

Excel for Mac 2011 and IBM SPSS Statistics
ver. 23.0. Normality of data was analyzed
using Kolmogorov-Smirnov and Shapiro-Wilk
analysis. The difference between two groups
was analyzed by using unpaired t test or Mann-
Whitney test (age, cell count, biochemistry
levels, number of spots, ADA) and chi square
(gender). Statistical significance defined as
p value <0.05. Diagnostic values such as
sensitivity, specificity, positive predictive value
(PPV), and negative predictive value (NPV) were
counted from 2x2 table. Positive MGIT culture
and/or ADA activity >40 U/L were used as the
gold standard of pleural TB.

Tika Adilistya Acta Med Indones-Indones J Intern Med

RESULTS

Of the 48 patients with presumed pleural TB,
with mean age of 49 years (range 19–80 years),
39 patients were confirmed as pleural TB based
on positive MGIT culture and/or ADA activity
>40 U/L as the gold standard. Characteristics of
both groups can be seen in Table 1. Based on the
gold standard, patients can be divided into two
groups: pleural TB (39 patients) and non pleural
TB (9 patients).

The median proportion of cell count in
pleural effusion fluid was 1530 cells/µL (range
80–64 680 cells/µL) in the group of patients with
confirmed pleural TB and 672.2 (SD 388.3) cells/
µL in the group of patients with no pleural TB.
The median proportion of MN cell count was
1203 cells/µL (range 64–58 212 cells/µL) in the
pleural TB group and 546 (SD 307.8) cells in
no pleural TB group. Both of cell count and MN
cell count were statistically different between
two groups.

Pleural fluid protein, pleural fluid LDH,
fluid/serum protein ratio, and fluid/serum LDH
ratio between two groups were not statistically
different.

Of the 39 patients in pleural TB group, all
of them (100%) were positive for T-SPOT.TB
(Table 1). The median spot of (Panel A – negative
control) was 38 (4–314) spots and (Panel B –
negative control) was 27 (0–328) spots. These

Table 1. Clinical and laboratory characteristics of the patients

Characteristics Pleural TB (n=39) Non pleural TB (n=9)

Age (years), mean (SD) 44.9 (16.5) 51.1 (18.4)

Gender (Male/Female) 18/21 7/2

Unilateral effusion on ultrasound 37 4

WBC count (/µL) 1530 (80 – 64.680)** 672.2 (388.3)*

MN count (/µL) 1203 (64 – 58.212)** 546 (307.8)*

Fluid protein (g/dL), mean (SD) 4.7 (0.95) 4.4 (0.71)

Protein ratio, mean (SD) 0.67 (0.11) 0.64 (0.07)

Fluid LDH (U/L) 613 (212 – 51,413)** 647.8 (395.25)*

LDH ratio, median (range) 1.24 (0.59 – 28.28) 0.79 (0.6 – 3.58)

T-SPOT.TB

- Panel A – negative control (spot) 38 (4 – 314)** -2 (1 – 12)**

- Panel B – negative control (spot) 27 (0 – 328)** 1.8 (4.7)*

ADA (U/L) 87.6 (48.9 – 461.5)** 17.95 (10.17)*

* mean (SD), ** median (range)

Tabel 2. 2x2 table of diagnostic test

Gold standard
Total

Positive Negative

T-SPOT.TB Positive 39 1 40

Negative 0 8 8

Total 39 9 48

results were statistically different with the results
in non pleural TB group with p-value <0.001.

The median levels of ADA were significantly
higher in pleural TB group than in non pleural
TB group as shown in Table 1.

From Table 2, sensitivity, specificity, PPV,
NPV, LR+, and LR- of T-SPOT.TB using PFMC
for the diagnosis of pleural TB are 100% (95% CI
90.97-100%), 88.89% (95% CI 51.75-99.72%),
97.5% (86.84-99.94%), 100% (63.06-100%), 9
(1.42 – 57.12), and 0, respectively.

DISCUSSION
Because TB disease can be difficult to

diagnose, IGRA such as T-SPOT.TB has recently
become popular as supportive diagnostic method
for TB. However, IGRA cannot distinguish
between active and latent TB infection (LTBI)
or healed TB if performed on peripheral blood
samples. In patients with pulmonary affection,
bronchoalveolar cells but not PBMCs showed
reactivity towards mycobacterial antigens. Thus,

Vol 48 • Number 1 • January 2016 Use of pleural fluid interferon-gamma enzyme-linked immunospot assay

it is suggestive that whilst only a small number
of antigen-specific T cells are found in peripheral
circulation, highly activated, antigen-specific
effector T cells accumulate at disease site and
rapidly produce Th-1-type cytokines.5

However, recently it was demonstrated that
enumeration of MTB-specific mononuclear cells
from the site of the infection by ELISPOT can
distinguish between active TB, LTBI, or other
disease with a high diagnostic sensitivity and
specificity. In smear-negative pulmonary TB,
the mean numbers of ESAT-6 and CFP10 spots
in lung mononuclear cells were 9.6- and 7.9-fold
higher than in PBMCs. In a smaller study on
pleural TB where only ESAT-6 antigen was used,
the mean number of ESAT-6 spots in PFMC was
15-fold higher than in PBMCs.8 In the present
study, the possibility of a reliable and rapid
diagnosis of pleural TB using a commercially
available MTB-specific ELISPOT in a routine
clinical practice was further evaluated.

Average age of subjects in this study was 49
years (range 19 – 80 years). It is similar with an
epidemiological analysis from the United States
with the mean age of 7549 patients was 49 years.
Epidemiologically, pleural TB predominates in
men with an overall male-to-female ratio of 2:1.9
In this study, the number of male subjects was
slightly more than the number of female subjects.

Of 48 patients, 39 patients (81.25%) were
diagnosed as having pleural TB, based on
positive MGIT culture and/or ADA activity more
than 40 U/L. The median/mean of pleural fluid
WBC count was 1530 cells/µL in pleural TB
group and 672 cells/µL in non pleural TB group
(Table 1). Both of them are statistically different.
In TB infection, cells, especially antigen-specific
T cells, are recruited and clonally expand at
the site of infection.5 This causes an increase
in WBC count. The median/mean of MN cells
count was 1203 cells/µL in pleural TB group
and 546 cells/µL in non pleural TB group, and
both groups are statistically significant. Although
neutrophils may be the predominant cells in the
pleural cavity in the initial stage, T lymphocytes
predominate thereafter. The compartmentalized
inflammatory process increases the permeability
of pleural capillaries. Along with impaired
lymphatic clearance due to parietal pleural

involvement, this leads to pleural fluid formation
and accumulation.9 Protein as inflammatory
products is accumulated in pleural fluid. A
pleural fluid protein concentration greater than
5 g/dL is found in 70% patients.9 The enzyme
lactat dehydrogenase (LDH) is found in the cells
of many body tissues, especiallly heart, liver,
lungs, brain, and skeletal muscle. When disease
or injury affects the cells containing LDH, the
cells lyse and LDH is spilled, causes an increase
in LDH activity in pleural TB effusion.10 In
this research, pleural fluid protein, fluid/serum
protein ratio, pleural fluid LDH, and fluid/
serum LDH ratio were not statistically different
between two groups. This is particularly because
in this research we only include patient with
exudative pleural fluid based on Light’s criteria.

The ELISPOT assay was positive in all 39
patients (100%) with pleural TB. In pleural
TB patients, the number of immunospots in
the pleural fluid ELISPOT assay was much
higher than the number of immunospots in
non pleural TB patients. The median value in
pleural TB group was 38 for the ESAT-6 antigen
(range 4–314) and 27 for the CFP10 antigen
(range 0–328) (Table 1). The median value of
immunospots in non pleural TB group was -2 for
ESAT-6 antigen (range 1–12) and 1.8 (SD 4.7)
spots for CFP10 antigen. Both groups showed
significantly difference with p-value of <0.001.
This is because the number of antigen-specific
T cells in pleural TB is higher and these cells
secrete more IFN-γ when stimulated.

A false-positive result of pleural fluid
ELISPOT was obtained for one patient, probably
due to the non-specific IFN-γ present in
the effusion before M. tuberculosis antigen
stimulation which trapped at the bottom of the
well immediately after the fluid was added.11
Furthermore, ESAT-6 and CFP10 are absent
from all BCG strains and from non-tuberculous
mycobacteria with the exception of M. kansasii,
M. szulgai, and M. marinum.12

In the diagnosis of pleural TB, ADA
activity is highly sensitive, simple, speed, and
relatively cheap. They cause the widespread
implementation and routine utilization of this
assay. In this research, ADA activity was higher
in pleural TB group than in non pleural TB group

Tika Adilistya Acta Med Indones-Indones J Intern Med

with p-value of <0.001 (Table 1). Increased ADA
activity in pleural TB is due largely to increased
activity of the ADA isoenzyme ADA-2, together
with the fact that only cells in which ADA-2
has been found are monocyte and macrophage.
Stimulation of those cells by live phagocytosed
microorganisms will cause cells to release
ADA.4,8

In this research, the sensitivity, specificity,
positive predictive value, and negative predictive
value of the IGRA ELISPOT assay using PFMC
for the diagnosis of pleural TB were 100%,
88.89%, 97.5%, and 100%, respectively.A
meta-analysis of 19 studies showed that pooled
sensitivity and pooled specificity for the pleural
fluid IGRA assay were 72% (95% CI, 55% - 84%)
and 78% (95% CI, 65% - 87%), respectively.13

As sensitivity of T-SPOT.TB was 100%,
theoretically this test is appropriate for screening
purpose. As specificity of T-SPOT.TB was only
88.89%, positive test results should always be
furtherly evaluated with caution because there
is a chance of 11.11% of false positive. As PPV
of T-SPOT.TB was 97.5%, if a patient with
presumptive pleural TB gets positive result, the
chance of this patient to be really sick is 97.5%.
As NPV of T-SPOT.TB was 100%, negative
result of patients with presumptive pleural TB
always ruled out a diagnosis of active pleural TB.

There are some limitations involved in the
use of IGRA in the diagnosis of pleural TB using
pleural fluid specimen. First, the T-SPOT.TB
assay is validated for PBMC sample therefore
the cut-off level for PBMC is well defined while
the cut-off level for PFMC sample has not yet
been defined. Several previous reports have used
various cut-off levels derived in heterogeneous
settings, this must be a problem in routine clinical
practice. A large, comparative study is required
to determine the appropriate cut-off levels for
positivity, and this may be different in areas with
a low or high incidence of TB. Second, pleural
TB could only be confirmed by TB culture in less
than half of cases (data not shown). However, we
have implemented a strict exclusion criteria. In
our clinical practice, considering that our country
has a high TB-burden setting, presumptive
pleural TB patients with exudative pleural fluid,
high ADA activity, with other common causes

of pleural effusion (liver failure, kidney failure,
malignancy, congestive heart failure) are already
excluded, will be given TB therapy although they
have negative TB culture.

CONCLUSION
The IGRA ELISPOT assay, T-SPOT.

TB,using pleural fluid samples showed a high
diagnostic accuracy for diagnosing pleural TB in
a high burden setting of TB infection. It suggests
that T-SPOT.TB assay may be an adjunctive but
useful method in the diagnosis of pleural TB.

ACKNOWLEDGMENTS
The research was financially supported by

CiptoMangunkusumo Hospital Research Grant
2015.

REFERENCES
1. Kang JY, Rhee CK, Kang NH, Kim JS, Yoon HK, Song

JS. Clinical utility of two interferon-gamma release
assays on pleural fluid for the diagnosis of tuberculous
pleurisy. Tuberc Respir Dis. 2012;73:143-50.

2. Liao M, Yang Q, Zhang J, et al. Gamma interferon
immunospot assay for pleural effusion mononuclear
cells for diagnosis of tuberculous pleurisy. Clin Vaccine
Immunol. 2014;21(3):347-53.

3. Reuter H. Pleural effusion and empyema in adult
tuberculosis. In: Schaaf HS, Zumla AI, Grange JM, et
al, eds. Tuberculosis, a comprehensive clinical reference.
1st ed. London: Elsevier Saunders; 2009. p. 342-50.

4. Abdallah AM, Gey van Pittius NC, Champion PA, et
al. Type VII secretion—mycobacteria show the way.
Nat Rev Microbiol. 2007;5(11):883-91.

5. Nemeth J, Winkler HM, Zwick RH, et al. Recruitment
of Mycobacterium tuberculosis specific CD4+ T
cells to the site of infection for diagnosis of active
tuberculosis. J Intern Med. 2009;265(1):163-8.

6. Light RW. Diagnostic principles in pleural disease. Eur
Respir J. 1997;10:476-81.

7. Light RW. Update on tuberculous pleural effusion.
Respirology. 2010;15:451-8.

8. Losi M, Bossink A, Codecasa L, et al. Use of a T-cell
interferon-gamma release assays for the diagnosis of
tuberculous pleurisy. EurRespir J. 2007;30(6):1173-9.

9. Porcel JM. Tuberculous pleural effusion. Lung.
2009;187:263-70.

10. Pagana KD, Pagana TJ. Mosby’s manual of diagnostic
and laboratory test. 5th ed. St. Louis: Elsevier Mosby;
2014. p. 329-30.

11. Lee LN, Chou CH, Wang JY, et al. Enzyme-linked
immunospot assay for interferon-gamma in the
diagnosis of tuberculous pleurisy. Clin Microbiol

Vol 48 • Number 1 • January 2016 Use of pleural fluid interferon-gamma enzyme-linked immunospot assay

Infect. 2009;15(2):173-9.
12. Anonymous. T-SPOT®.TB, an aid in the diagnosis of

tuberculosis infection, 96-well plate format. [package
insert]. Oxon: Oxford Immunotec; 2007.

13. Aggarwal AN, Agarwal R, Gupta D, Dhooria S, Behera
D. Interferon gamma release assays for diagnosis of
pleural tuberculosis: a systematic review and meta-
analysis. J Clin Microbiol. 2015;53(8):2451-9.