Tuberculous Pleural Effusion in

Children*
Jose M. Merino, MD; Isabel Carpintero, MD; Teresa Alvarez, MD;
Jesus Rodrigo, MD; Jesus Sanchez, MD; and Jose M. Coello, MD

Study objectives: To describe the age distribution, clinical, laboratory, radiographic, and
bacteriologic findings of pediatric patients with tuberculous pleural effusion.
Design: A retrospective study.
Patients and methods: We have identified all cases of primary pulmonary tuberculosis in children
< 18 years, reported to the health department. We have collected information from medical
records regarding demographics, clinical findings, bacteriologic results, and evolution. Chest
radiographs obtained at the time of initial evaluation were reviewed independently by two groups
of radiologists who were blind to the clinical and epidemiologic data.
Results: Between January 1983 and December 1996, 175 children <18 years were diagnosed as
having primary pulmonary tuberculosis. Among them, 39 patients (22.1%) showed pleural
effusion on chest radiograph. The mean age of patients with tuberculous pleural effusion was
significantly higher (13.52 ! 0.5 years vs 6.97 ! 0.42 years). The sensitivity of the tuberculin test
is 97.4% for an induration > 5 mm. Pleural fluid analysis shows a lymphocytic exudative effusion.
Chest radiograph review showed unilateral pleural effusion in all cases. Pleural effusion was the
sole radiographic manifestation in 41% of cases. Parenchymal disease is associated in 23 cases
(59%). Bacteriologic confirmation of tuberculosis was achieved in 22 cases (56.4%). Cultures of
pleural fluid and biopsy material both yielded Mycobacterium tuberculosis in 15 of 34 (44.1%)
and 12 of 18 (66.6%), respectively, for samples under study. Pleural biopsy specimens showed
granulomatous inflammation in 18 of 23 cases (78.3%). Antituberculous therapy for 6 to 9 months
was effective in all cases. Transient side effects occurred in 1 of 39 patients (2.9%).
Conclusions: Pleural effusion accounts for 22.1% of cases of pediatric pulmonary tuberculosis.
Parenchymal consolidation is the most common associated radiographic finding. Bacteriologic
confirmation was achieved in 56.4% of cases. A short course of chemotherapy is effective.
(CHEST 1999; 115:26 30)
Key words: bacteriology; children; clinical outcome; pleural effusion; radiology; tuberculosis
Abbreviations: ADA ! adenosine deaminase activity; LDH ! lactate dehydrogenase; TB ! tuberculosis; TPE !
tuberculous pleural effusion

effusion due to Mycobacterium tuberculoP sisleuralis categorized

as extrapulmonary disease.
1,2

Pleurisy with effusion develops as a complication of

primary pulmonary tuberculosis (TB) in 2 to 38% of
children with pulmonary disease.27 Effusion is not a
common feature of primary pulmonary TB in young
children and it is more likely to be observed in
adolescents and adults.3,7
Establishing a diagnosis of tuberculous pleural
effusion (TPE) can be difficult because the classic
findings (lymphocytic exudative pleural effusion,
*From the Pediatric (Drs. Merino, Carpintero, Alvarez, Rodrigo,
and Sanchez) and Radiologic (Dr. Coello) Departments,
General Yague Hospital, Burgos, Spain.
Manuscript received February 6, 1998; revision accepted August
27, 1998.
Correspondence to: Jose M. Merino, MD, Pediatric Department,
Hospital General Yague, Avda. del Cid, s/n, 09006 Burgos, Spain
26

Downloaded From: http://journal.publications.chestnet.org/ on 02/05/2015

pleural granulomata, and cutaneous sensitivity to

purified protein derivative [PPD]) have low specificity and sputum, pleural fluid, and pleural biopsy
cultures have a low rate of success.
We have retrospectively reviewed our experience
from 1983 to 1996 in pediatric patients with TPE to
determine the following: (1) the age distribution of
pediatric patients with TPE; (2) the main clinical and
laboratory findings; (3) the associated radiologic
patterns; (4) the efficacy of mycobacterial cultures in
the diagnosis of pleural TB; and (5) the outcome of
patients.
Materials and Methods
We selected for retrospective review all cases of primary
pulmonary TB in children " 18 years reported to the health
Clinical Investigations

department from January 1983 to December 1996. All patients

were admitted to our hospital, a tertiary care center with a
reference population of 294,763 people, 68,155 of them " 18
years. We defined a case of TB if (1) M tuberculosis was detected
from a clinical specimen, or (2) a child had clinical evidence of
current disease and any of the following: (a) a history of contact
with an adult case of TB; (b) positive tuberculin skin test (! 5
mm); (c) suggestive appearances on chest radiograph; and (d)
favorable response to specific antituberculous therapy. Clinical,
laboratory, bacteriologic, radiologic, and treatment data were
reviewed.
TPE was diagnosed if the patient showed a chest radiograph
interpreted by a radiologist as depicting a pleural effusion and at
least one of the following criteria: (1) positive culture for M
tuberculosis from sputum, gastric aspirate, pleural fluid, or
pleural biopsy specimen; (2) acid-fast bacilli in sputum, gastric
aspirate, pleural fluid, or biopsy tissue; (3) pleural tissue histopathology compatible with TB (caseating granulomas with Langhans giant cells, epithelioid cells, and lymphocytes); (4) compatible clinical picture with positive tuberculin test and one of the
following: lymphocytic pleural fluid (# 50%), exudative fluid
(protein # 3 g/dL or lactate dehydrogenase [LDH] # 200 U/L)
or pleural fluid levels of adenosine deaminase activity (ADA)
# 40 U/L; and (5) radiographic pleural effusion that resolved
with appropriate antimycobacterial therapy.
Tuberculin skin testing was made by means of an intradermal
injection of 2 tuberculin units of PPD. All skin tests were
undertaken by nursing personnel who are familiar with the
technique. Induration was measured in 48 to 72 h and recorded
in millimeters. A positive test was considered if the palpable
induration was ! 5 mm.
Chest radiographs obtained at the time of initial evaluation
were reviewed independently by two groups of radiologists who
were blind to the clinical and epidemiologic data. Chest radiograph interpretation was made on the basis of the following
patterns: (1) parenchymal consolidation (unilobar, multilobar,
segmental, masslike); (2) atelectasis (lobar, segmental); (3)
lymphadenopathy (hilar, paratracheal, hilar prominence, tracheal
compression); (4) pleural effusion (effusion, empyema); and (5)
miliary TB.
Specimens of sputum, gastric washing, pleural fluid, and biopsy
tissue were obtained from patients with pleural effusion and
studied for acid-fast bacilli detection by Ziehl-Neelsen smear and
microscopic evaluation. Mycobacterial cultures were performed
using Lowenstein-Jensen medium.
The statistical analyses were performed using the MannWhitney and $2 tests.

Table 1Epidemiologic and Tuberculin Test Data

Pleural
Effusion

Cases
Sex
M
F
Mean age, yr, mean
05
610
# 10
Tuberculin test, mm
Induration, mm,
mean % SEM
"5
59
1014
1519
! 20

No.

No.

39

22.2

136

77.8

Between January 1983 and December 1996, 175

children " 18 years were diagnosed as having primary pulmonary TB. Chest radiograph review disclosed 39 cases of TPE, which represented 22.1% of
all forms of pulmonary disease due to M tuberculosis.
Epidemiologic, demographic, and skin test data
are shown in Table 1. A history of contact with a
tuberculous case was present in 45 of 175 patients
(25.7%), usually immediate family members
(75.5%). Table 2 shows the main clinical and laboratory findings.
Radiologic records were available for review in all
patients. Associated chest radiograph findings in

p Value*

22
56.4
17
43.5
13.52 % 0.5
1
2.5
5
12.8
33
84.6

78
57.3
58
42.6
6.97 % 0.42
70
51.4
30
22.0
36
26.4

NS
NS
" 0.001

13.76 % 0.52

14.2 % 0.73

NS

1
2
17
15
4

2.5
5.1
43.5
38.4
10.2

11
11
37
55
22

8.0
8.0
27.2
40.4
16.1

*NS ! not significant.

patients with pleural effusion are shown in Table 3.

Parenchymal consolidation is the most common associated radiographic finding (59%). No statistical
differences in the prevalence of parenchymal disease
between children aged 10 years or more and younger
were found (p ! 0.76). There is a good agreement
between the two groups of radiologists in the evaluation of chest radiographs (kappa index of 0.62).
Bacteriologic confirmation of TB was achieved in
22 cases (56.4%). A total of 91 clinical specimens (24
sputum, 12 gastric washing, 32 pleural fluid, 2
synovial fluid, 3 ascitic fluid, and 18 pleural biopsy

Table 2Clinical and Laboratory Findings*

Pleural
Effusion
No.

Results

No Pleural
Effusion

Clinical data
Malaise
6
15.4
Anorexia
12
30.8
Fatigue
14
35.9
Respiratory distress
12
30.8
Thoracic pain
31
79.5
Cough
21
53.8
Weight loss
5
12.8
Mean temperature, C 38.06 % 0.13
Laboratory data, mean % SEM
WBC, &109/L
7.6 % 0.3
Neutrophils, %
63.06 % 2.02
LDH, U/L
394.4 % 25.8
CPR, mg/L
75.53 % 18
ESR, mm/h
48.88 % 3.08

No Pleural
Effusion
No.

p
Value

21
15.4
NS
43
31.6
NS
28
20.6 " 0.05
12
8.8 " 0.001
17
12.5 " 0.001
70
51.5
NS
19
14.0
NS
38.31 % 0.08
NS
11.5 % 0.17 " 0.001
60.37 % 1.37
NS
470.76 % 17.6 " 0.05
36.3 % 4.3
" 0.05
40.14 % 2.3
" 0.05

*NS is expanded in Table 1 footnote. CPR ! C-reactive protein;

ESR ! erythrocyte sedimentation rate.
CHEST / 115 / 1 / JANUARY, 1999

Downloaded From: http://journal.publications.chestnet.org/ on 02/05/2015

27

Table 3Chest Radiograph Findings in Patients With

Pleural Effusion
Radiographic Features

No.

Total cases
Unilateral pleural effusion
Right side
Left side
Pleural effusion only
Pleural effusion and parenchymal disease
Parenchymal consolidation
Lobar
Multilobar
Segmental
Upper lobe
Middle lobe
Lower lobe
Bulging fissures
Atelectasis
Mediastinal lymphadenopathy
Hilar only
Pneumothorax
Miliary TB
Cavitation

39
39
26
13
16
23
17
6
8
3
1
9
15
17
4
9
7
2
2
1

100
100
66.6
33.3
41.0
59.0
43.5

43.5
10.2
23.0
5.1
5.1
2.5

samples) from 37 patients were examined by microscopy. All but 5 of 18 (27.7%) pleural biopsy specimens were negative on microscopy examination.
Ziehl-Neelsen stain recovery rate was 5.49%. A total
of 70 clinical specimens (4 sputum, 9 gastric washing,
34 pleural fluid, 3 ascitic fluid, 2 synovial fluid, and
18 pleural biopsy samples) from 35 patients were
submitted for Lowenstein-Jensen culture. M tuberculosis was recovered from 34 of 70 (48.5%) clinical
specimens. Pleural fluid and biopsy material cultures
yielded M tuberculosis in 15 of 34 (44.1%) and 12 of
18 (66.6%), respectively, of samples studied. The
median time for growth detection was 34 and 35 days
for pleural biopsy and pleural fluid LowensteinJensen cultures, respectively. No statistical differences in mean ADA levels (77.4 % 15.8 U/L vs
63 % 7.0 U/L, p ! 0.61), mean tuberculin skin test
induration (12.95 % 0.8 mm vs 15.5 % 0.9 mm,
p ! 0.06), and parenchymal involvement (p ! 0.09)
were found between the positive and negative bacteriology group. However, mean age is significantly
higher in bacteriology-positive patients (14.2 % 0.45
vs 12.1 % 0.99, p ! 0.04). Table 4 shows the cytologic and biochemical analysis of pleural fluid.
Histologic findings from pleural biopsy specimens
were available in 23 patients and 18 of these 23
(78.3%) showed granulomatous inflammation. Of
them, nine (50%) were also culture positive. Four
patients (22.2%) showed granulomatous inflammation in pleural biopsy histologic examination but
negative Ziehl-Neelsen stains and mycobacterial cultures.
Treatment changed over the study period. From
28

Downloaded From: http://journal.publications.chestnet.org/ on 02/05/2015

Table 4 Pleural Fluid Analysis

Laboratory Data
Cells/'L
" 2,000
2,0014,000
4,0016,000
# 6,000
Lymphocytes, %
# 50%
Proteins, g/dL
" 4.0
4.05.0
5.16.0
# 6.0
Glucose, g/dL
LDH, U/L
" 500
5011,000
1,0011,500
1,5012,000
# 2,000
LDHpf/LDHp*
" 1.0
1.12.0
2.13.0
3.14.0
# 4.0
ADA, U/L
# 40
" 40
4060
6180
# 80
pH

No.
32
9
13
9
1
32
23
33
2
18
12
1
33
30
5
14
5
3
3
23
3
8
3
4
5
20
18 (90%)
2
8
5
5
33

Mean % SEM
3,186 % 291

71.78 % 4.3
4.96 % 0.098

69.12 % 3.02
1,009 % 116.3

3.21 % 0.6

73.8 % 11.9

7.25 % 0.037

*LDHpf/LDHp ! LDH (pleural fluid)/LDH (plasma).

1983 to 19921993, most of the patients received a

combination of isoniazid, rifampicin, and ethambutol
for 9 months (27 of 39 [69.4%] patients with TPE in
our review). After that, patients were treated with a
combination of isoniazid, rifampicin, and pyrazinamide for 6 months (8 of 39 [20.5%] patients with
TPE in our study). A short course of prednisone
therapy (1 month) was added in nine patients (23%)
if a large-volume pleural effusion or a prolonged
course of fever was present. There was a complete
resolution of the effusion in 34 patients (87.1%).
Minimal residual pleural thickening was seen in five
patients (13%) 1 year after the diagnosis. One patient
(2.5%) showed transient elevated results of liver
function tests.
Discussion
Published reports describe TPE as a complication
of primary pulmonary TB in 2 to 38% of children
with pulmonary disease.1 4,6,7 Effusion is the sole
radiographic manifestation of primary pulmonary TB
Clinical Investigations

in 38 to 63% of cases.3 In our series, TPE accounts

for 39 of 175 cases (22.1%) of primary pulmonary TB
and in 16 patients (9.6% of children with primary
pulmonary TB), pleural effusion is the sole radiographic manifestation of primary pulmonary TB.
TPE is very uncommon in young children and is
more likely to be observed in adolescents and
adults.35,7 Mean age of patients with TPE was
higher (13.52 years) than children without pleural
effusion (6.97 years) in our study. Only six patients
(15.4%) with TPE are " 10 years of age. However,
no statistical differences in the prevalence of parenchymal abnormalities between children aged 10
years or more and younger were found in our review
(p ! 0.76).
TPEs are usually unilateral.2,3,7 No cases of bilateral effusion were seen in our study. Pleural effusion
accompanied by underlying parenchymal disease was
seen in 59% of our patients1,3,7 (Table 3). Parenchymal consolidation and bulging fissures are the most
common associated radiographic manifestations.
The diagnosis of childhood TB is based on clinical
findings, chest radiography, tuberculin skin testing, and
a history of close contact with an adult case of pulmonary TB.4 A history of contact with a tuberculous case
was present in 25.7% of cases in our study. The clinical
picture of childhood TB is not specific.4 Patients with
pleural effusion frequently had chest pain and respiratory distress.2 No differences in other clinical signs are
detected in our review (Table 2).
Tuberculin tests showed a high sensitivity in our
patients. Thirty-eight patients (97.4%) showed an
induration of ! 5 mm. Significant cutaneous reactivity to tuberculin among patients with TPE has been
reported in the literature.1 However, no differences
in mean tuberculin skin test induration between the
two groups were observed (Table 1).
Analysis of effusion usually shows a lymphocytic
exudative effusion.1 In our series, 23 of 32 patients
(71%) showed lymphocytic (# 50%) fluid. A neutrophil predominance (63.6 % 6.3%, mean % SEM)
was found in the other nine patients (29%).
Some reports have shown the importance of the
determination of pleural ADA in the differential
diagnosis of tuberculous and nontuberculous pleural
effusions.8 12 Specificity and sensitivity of the ADA
test in TB are very high.8 10 Mean ADA activity was
73.8 % 11.9 (mean % SEM) in our series and 2 of 20
patients (10%) showed ADA activity " 40 U/L (Table 4). The sensitivity of test for values of ! 40 U/L
is 90%. No statistical differences in ADA levels
between negative or positive bacteriology and negative or positive histologic pleural biopsy findings
were found.
Bacteriologic confirmation of childhood TB
ranged from 20 to 42%4,6 in reported series. In our

series, bacteriologic confirmation was achieved in 22

patients (56.4%). Most of our cases (17/22) were
diagnosed on the basis of a culture of M tuberculosis
from pleural biopsy material or pleural fluid.
Ziehl-Neelsen stain recovery rate is very low in our
review (5.49%), as in reported series.4,6 Gastric
aspirate, pleural fluid, and biopsy material cultures
yielded M tuberculosis in 2 of 9 (22.2%), 15 of 34
(44.1%), and 12 of 18 (66.6%), respectively, of
samples submitted for culture. No clinical, laboratory, and radiographic differences between patients
with and without positive pleural biopsy culture were
seen.
Pleural biopsy histologic examination revealing
granulomatous inflammation is frequently used as a
diagnostic criterion for pleural TB.1 In our series, 18
of 23 (78.3%) pleural biopsy specimens showed
granulomatous inflammation. Four patients showed
positive pleural biopsy histologic findings but negative mycobacterial cultures and Ziehl-Neelsen stains.
Pleural fluid LDH levels were significantly higher in
patients with positive histologic findings (1,171 % 166
U/L vs 514 % 97 U/L, p ! 0.003). The combination of
culture and histologic examination of pleura has been
described as the most sensitive diagnostic test for
pleural TB. Our data support the use of pleural fluid
and biopsy cultures and pleural histologic examination
in the approach to the pediatric patient with suspected
TPE.
Tuberculous pleurisy is associated with low bacterial populations, as shown by the high index of
negative sputum smears and the lack of positive
cultures of pleural fluid and biopsy material.13 A
short course of chemotherapy with two bactericidal
drugs is effective in the treatment of tuberculous
pleurisy.1315 In our series, antituberculous therapy
for 6 to 9 months with rifampicin and isoniazid, along
with initial supplement of ethambutol/pyrazinamide,
was effective in all cases. Side effects of the drugs
occurred in one case (transient toxic hepatitis),
which represents 2.9% of all cases.
References
1 Seibert AF, Haynes J Jr, Middleton R, et al. Tuberculous
pleural effusion: twenty-year experience. Chest 1991; 99:883
886
2 Waagner DC. The clinical presentation of tuberculous disease in children. Pediatr Ann 1993; 22:622 628
3 Page McAdams H, Erasmus J, Winter JA. Radiologic manifestations of pulmonary tuberculosis. Radiol Clin North Am
1995; 33:655 676
4 Shaaf HS, Beyers N, Gie RP, et al. Respiratory tuberculosis in
childhood: the diagnostic value of clinical features and special
investigations. Pediatr Infect Dis J 1995; 14:189 194
5 Tournier G. Tuberculose de lenfant. Rev Pneumol Clin 1994;
50:288 294
6 Pineda PR, Leung A, Muller NL, et al. Intrathoracic paediCHEST / 115 / 1 / JANUARY, 1999

Downloaded From: http://journal.publications.chestnet.org/ on 02/05/2015

29

7
8

9
10
11

atric tuberculosis: a report of 202 cases. Tuber Lung Dis

1993; 74:261266
Agrons GA, Markowitz RI, Kramer SS. Pulmonary tuberculosis in children. Semin Roentgenol 1993; 2:158 172
Fontan J, Verea H, Perez J, et al. Diagnostic value of
simultaneous determination of pleural adenosine deaminase
and pleural lysozyme/serum lysozyme ratio in pleural effusions. Chest 1988; 93:303307
Ocana I, Martnez-Vazquez JM, Segura RM, et al. Adenosine
deaminase in pleural fluids: test for diagnosis of tuberculous
pleural effusion. Chest 1983; 84:5153
Orriols R, Coloma R, Ferrer J, et al. Adenosine deaminase in
tuberculous pleural effusion. Chest 1994; 106:16331634
Bandres R, Abal J, Blanco J, et al. Actividad de adenosinde-

30

Downloaded From: http://journal.publications.chestnet.org/ on 02/05/2015

12
13
14
15

saminasa en lquido pleural: estudio realizado en 64 casos.

Arch Bronconeumol 1994; 30:8 10
Serra J, Jane X, Sole C, et al. Adenosina desaminasa como
parametro diagnostico del derrame pleural tuberculoso. Med
Clin 1991; 96:636 637
Canete C, Galarza I, Granados A, et al. Tuberculous pleural
effusion: experience with 6 months of treatment with isoniazid and rifampicin. Thorax 1994; 49:1160 1201
Dutt AK, Moers D, Stead WW. Tuberculous pleural effusion:
6-month therapy with isoniazid and rifampin. Am Rev Respir
Dis 1992; 145:1429 1432
Malik SK, Behera D, Gilhotra R. Tuberculous pleural effusion and lymphadenitis treated with rifampin-containing regimen. Chest 1987; 92:904 905

Clinical Investigations