Вы находитесь на странице: 1из 4

ORIGINAL STUDIES

Childhood Pleural Tuberculosis


A Review of 45 Cases
Andrea T. Cruz, MD, Lydia T. Ong, PA, and Jeffrey R. Starke, MD

of pleural TB often is based on the triad of a positive tuberculin skin


Background: Pleural tuberculosis (TB) can be seen in isolation or com-
test (TST), close contact with a known case of contagious TB
plicate untreated pulmonary disease in children. It is infrequently suspected
(called a source case), and compatible clinical and radiographic
in low incidence countries, leading to delays in diagnosis and treatment.
findings. This study describes one US center’s recent experience
Methods: This was a retrospective descriptive case series of children ⱕ18
with a consecutive series of children with tuberculous pleural
years old with pleural TB seen in a referral clinic in Houston, Texas from
effusions in terms of epidemiology, clinical manifestations, diag-
1984 to 2008. Medical record data, radiographs, and health department
nosis, treatment, and outcomes.
records were reviewed. Epidemiologic, clinical, radiographic, and micro-
biologic findings are described.
Results: Forty-five children (mean age 10.9 years) were diagnosed with TB. METHODS
Contacts with cases of infectious tuberculosis were identified for 20 children This retrospective descriptive case series identified pediatric
(44%). Positive tuberculin skin tests were seen in 89%. Seventy-three percent patients ⱕ18 years old with TB seen at the Ben Taub General
of children had concomitant pulmonary parenchymal disease or intrathoracic Hospital Children’s Tuberculosis Clinic from 1984 to 2008. This is
lymphadenopathy. Cultures were attempted in 40 of 45 children (89%). No 1 of 2 pediatric TB clinics serving greater Houston, Texas and a
patient had acid-fast bacillus smear-positive pleural fluid, but pleural fluid multistate catchment region. The clinic has seen more than 8000
cultures were positive in 56%. Cultures from other respiratory sources were patients for TB exposure, infection, or disease, including more
positive in 48% of patients. Although all patients had symptomatic resolution than 860 patients with TB disease in a 25-year period. All patients
and radiographic improvement, chest radiographs were improved but abnor- were cared for by at least one of the authors (J.R.S.).
mal in 62% of patients at the end of therapy. No child developed relapse or TB was defined as a pleural effusion in a patient with at least
recrudescence after therapy was discontinued. one of the following: (i) epidemiological risk factors for tuberculosis,
Conclusions: Pleural TB was usually associated with a positive tuberculin (ii) positive TST (ⱖ5 mm), (iii) positive acid-fast smear or culture for
skin test, and lung parenchymal infiltrates were seen commonly. All M. tuberculosis, or (iv) histopathologic findings compatible with TB
children had symptomatic resolution and radiographic improvement, and (eg, caseating granulomas), and the exclusion of reasonable alterna-
6-month courses of therapy resulted in disease resolution for children with tive diagnoses. Chest radiographs obtained at the initiation of therapy
isolated pleural or pleural and pulmonary parenchymal disease. were evaluated for pleural effusion, parenchymal consolidation, in-
trathoracic adenopathy, and miliary disease.
Key Words: tuberculosis, pleural, childhood Data on demographic characteristics, cultures, radiographs,
(Pediatr Infect Dis J 2009;28: 981–984) treatment regimens, and prognosis were collected prospectively
and were abstracted from medical records retrospectively. Addi-
tional data were obtained, with verbal permission, from local
health department databases. Data were analyzed with Epi Info
version 3.4.3 (Atlanta, GA). Before initiation, the Baylor College
P leural tuberculosis (TB) typically has been considered a disease
of adulthood, estimated to comprise approximately 4% of
disease cases.1 However, tuberculous pleural effusions can com-
of Medicine institutional review board approved the study.

plicate 12% to 38% of cases of children with untreated pulmonary RESULTS


TB.2– 4 Childhood pleural TB has been considered a disease of Forty-five children had TB during a 25-year period, ac-
adolescence, and the mean age at diagnosis is 13 years.4 Tuber- counting for 5.3% of the 860 children with TB disease seen during
culous pleural effusions are typically unilateral, and are associated this time. Twenty-two children (48.9%) were male. The mean and
with pulmonary parenchymal disease in 40% to 50% of children.4 median ages were 10.9 years and 10.8 years, respectively (range:
The diagnosis of pleural TB has been limited by several 16 months to 17 years). Twenty-four (53.3%) were Hispanic, 14
factors. First, the paucibacillary nature of pleural TB reduces the (31.1%) non-Hispanic black, 4 (8.9%) Asian, and 3 (6.7%) non-
yield of positive stain and culture for acid-fast bacillus. Second, Hispanic white. Thirty-three percent of children and 64% of
there is relatively less experience with newer diagnostic modalities parents were foreign-born, with the most common foreign country
such as polymerase chain reaction (PCR) and adenosine deaminase in of origin being Mexico. The diagnosis of tuberculous pleural
children compared with adults, though both assays have been used to effusion was based on a combination of radiographic findings
diagnose childhood pleural TB. Third, there may be a low index of (100%), epidemiologic risk factors (96%), positive TST (89%),
suspicion for tuberculosis in some parts of the industrialized world positive culture (44%), histopathology (27%), and PCR (7%)
when a child presents with an isolated pleural effusion. The diagnosis (Table 1). Of the 3 children (11%) with negative TSTs, 1 each had
miliary disease, meningitis, and pericarditis. All children with
isolated pleural effusions had positive TSTs.
Accepted for publication April 2, 2009.
From the Department of Pediatrics, Baylor College of Medicine, Houston, TX. Symptoms
Address for correspondence: Andrea T. Cruz, MD, 6621 Fannin Street, Suite The most common symptoms were cough (80%), fever (67%),
A210, MC 1-1481, Houston, TX 77030. E-mail: acruz@bcm.edu.
Copyright © 2009 by Lippincott Williams & Wilkins
pleuritic chest pain (58%), and weight loss (29%); less frequently seen
ISSN: 0891-3668/09/2811-0981 were hemoptysis (9%) and night sweats (7%). The child with TB
DOI: 10.1097/INF.0b013e3181a8568b meningitis and pleural TB presented with headache, vomiting, and

The Pediatric Infectious Disease Journal • Volume 28, Number 11, November 2009 www.pidj.com | 981
Cruz et al The Pediatric Infectious Disease Journal • Volume 28, Number 11, November 2009

TABLE 1. Number (%) of Patients Fulfilling Each of the Diagnostic Criteria

Epidemiologic Risk Positive AFB No. Children


Positive TST Conclusive Histopathology
Factors Present Culture or Smear Meeting Criteria (%)

⫹ ⫹ ⫹ ⫹ 4 (8.9)
⫹ ⫹ ⫹ 12 (26.7)
⫹ ⫹ ⫹ 5 (11.1)
⫹ ⫹ 19 (42.2)
⫹ ⫹ ⫹ 1 (2.2)
⫹ ⫹ 2 (4.4)
⫹ ⫹ 1 (2.2)
⫹ ⫹ 1 (2.2)
Totals 40 (88.9%) 43 (95.6%) 20 (44.4%) 12 (26.7%) 100%
AFB indicates acid-fast bacilli.

identified in 40 children (89%) (Table 2): close contact with


TABLE 2. Historical and Microbiologic Findings
individuals from countries with high prevalence of TB (30);
Associated With Childhood Pleural Tuberculosis in a contact with known suspect TB case at the time of diagnosis (8),
Series of 45 Patients contact with HIV-infected adult (1), and contact with an adult with
a history of incarceration (1). The only child with a known
Findings No. Patients (%)
underlying immunocompromising medical condition was a 5-year-
Known source case before diagnosis 8 (17.8%) old boy with acute lymphoblastic leukemia.
Other epidemiologic risk factors 35 (77.8%)
TST Positive 40/45 (88.9%) Diagnostic Evaluation
AFB smear positive
Total 3/40 (7.5%)
Thirty-seven children were hospitalized for pleural TB. The
Pleural fluid 0/9 8 children who were treated entirely in the outpatient setting were
Other sites 3/31 (9.7%) asymptomatic or mildly symptomatic, had small effusions, and/or
M. tuberculosis culture-positive by site had known source cases from whom susceptibility data were
Total 20/40 (50%)
Pleural fluid 5/9 (55.6%)
available. Reasons for hospitalization included the need for surgi-
Nonpleural fluid 15/31 (48.4%) cal procedures (thoracentesis, video-assisted thoracoscopic surgery
Pleural biopsy 8/16 (50%) (VATS), pericardiocentesis) (28), collection of gastric aspirates
Gastric aspirate 4/8 (50%) (8), and altered mentation (1). Mean duration of hospitalization
Sputum 3/4 (75%)
Pericardial fluid 0/2
was 7.5 days (median 5 days), with range of 3 days to 7.5 weeks.
Peritoneal lymph node 0/1 The children with the longest hospitalizations were those with
PCR–pleural fluid 3/5 (60%) concomitant TB pericarditis; both required pericardiocenteses and
Histopathology of pleural tissue* pericardial windows, but did not require pericardiectomy. A
AFB smear positive 4/16 (25%)
Caseating granulomas 8/16 (50%)
3-year-old boy with endobronchial TB required 2 bronchoscopies.
Necrosis 1/16 (6.3%) The mean duration of hospitalization for children requiring VATS
Nonspecific inflammation 5/16 (31.3%) or thoracentesis was 9.5 days.
Nondiagnostic 3/16 (18.8%)
*Some patients had ⬎1 histopathological feature noted. Twelve of 16 children had Radiographic Studies
either positive AFB smears or caseating granulomas noted. Twelve patients (26.7%) had isolated pleural TB, 32
AFB indicates acid-fast bacilli. (71.1%) had pulmonary parenchymal involvement (infiltrates
and/or hilar or mediastinal lymphadenopathy), and 1 had miliary
TB. The most common radiographic abnormalities are described in
altered mentation. The 2 children with concomitant pericardial TB Table 3. Of the children with pulmonary parenchymal involve-
presented with chest pain and shortness of breath exacerbated by ment, 5 also had extrapulmonary disease (2 peritoneal, 2 pericar-
reclining, and the adolescent with peritoneal and pleural TB presented ditis, 1 meningitis); 1 child had endobronchial tuberculosis in
with a 2 to 3 month history of abdominal pain. Mean symptom association with pleural TB.
duration before diagnosis was 4.4 weeks (median 3 weeks; range 3
days to 4 months). Six children (13%) were asymptomatic and were Microbiologic Studies
diagnosed after being identified via contact investigations. Of the 12 Cultures were positive for M. tuberculosis in 20 children.
children with isolated pleural TB, 10 were symptomatic at diagnosis. Cultures were positive from the adult source case alone in 15
The most common symptoms were fever (8/12, 67%), weight loss (5, cases, the child alone in 15 episodes, both the child and adult
42%), and pleuritic chest pain (4, 33%). These symptoms were not source case in 5 instances, and no culture results were positive for
statistically significantly different from children with tuberculous either the child or a source case in 10 cases (Table 1). Pleural fluid
effusions with concomitant intrathoracic disease. was obtained in 9 children and TB was isolated in 5 cases (56%).
Pleural fluid cell count was obtained in 6 patients, with a mean of
Epidemiology 3321 white blood cells/mm3 (range: 37– 6957 white blood cells/
Source cases were identified for 20 children (44.4%), and in mm3) with a lymphocytic predominance and 5821 red blood
12 of 20 cases (60%), the adult source case was identified only cells/mm3 (range: 590 –14300 red blood cells/mm3). Gastric aspi-
after the child was diagnosed. Source cases were most commonly rate or sputum cultures were positive in one-third of children who
an aunt/uncle (8), grandparent (5), or parent (4). The source case had apparently isolated pleural disease by chest radiography.
was a household contact for 17 of 20 children. Only 2 children had Histopathology of pleural tissue most commonly showed case-
nonrelative source cases. Epidemiological risk factors for TB were ating granulomas (50%) and acid-fast organisms (25%). Six

982 | www.pidj.com © 2009 Lippincott Williams & Wilkins


The Pediatric Infectious Disease Journal • Volume 28, Number 11, November 2009 Childhood Pleural TB

DISCUSSION
TABLE 3. Radiographic Findings Seen in 45 Children
Several differences were noted between this study popula-
With Pleural Tuberculosis
tion and previously reported data in terms of prevalence, age
No. Children Abnormalities distribution, radiographic findings, and culture yield. Pleural TB
Finding With Present at End was less commonly seen in this patient population than has been
Finding (%)* of Therapy (%)* previously reported in other pediatric series: 5% versus 22% of all
All cases
cases of TB disease.4 This might reflect our referral population,
Total 45/45 (100%) 28/45 (62.2%) which is heavily skewed toward preschool-aged children identified
Right 24 (53.3%) 14/24 (58.3%) in the course of health department contact investigations. Many of
Left 18 (40%) 11/18 (61.1%) the children diagnosed with TB disease in our clinic are asymp-
Bilateral 3 (6.7%) 3/3 (100%)
Isolated pleural tomatic (often with isolated hilar or mediastinal lymphadenopathy)
Total 12 5/12 (41.7%) or are identified early in their clinical course as a result of active
Right 7 2/7 (28.6%) surveillance. Whereas pleural TB has been more frequently re-
Left 4 2/4 (50%) ported in adolescents,3,4 in this series 7 of 45 (15.6%) of children
Bilateral 1 1/1 (100%)
Effusion ⫹ adenopathy were younger than 5 years of age and 28 of 45 (62.2%) were
Isolated effusion ⫹ adenopathy 5 3/5 (60%) prepubertal. This finding also may be a phenomenon of our
Parenchymal

involvement relatively young patient population. Most of the children diag-
Total 27 19/27 (70.4%)
Right upper lobe 4 3/4 (75%)
nosed with TB disease in our clinic population are children in the
Right middle lobe 9 5/9 (55.6%) first 5 years of life. Although a relatively smaller proportion of
Right lower lobe 11 8/11 (72.7%) these young children might have TB disease, the absolute number
Left upper lobe 6 5/6 (83.3%) of pleural TB cases is greatest in the youngest children.
Left lower lobe 8 6/8 (75%)
Cavity 3 3/3 (100%)
Chest radiographic findings showed parenchymal involve-
Pericardial effusion 2 0 ment in more children (73%) than have been reported in prior
Miliary disease 1 0 studies (56%).4 Few children had bilateral pleural effusions. Ef-
*The numbers do not sum to 45 because some children had ⬎1 radiographic abnor- fusions were somewhat more common on the right, perhaps
mality.

reflecting intrathoracic lymphatic drainage patterns. The majority
Eleven children (24%) had multi-lobar involvement with TB. of children continued to have abnormal chest radiographs at the
end of therapy, though no child had recurrence of disease after
termination of therapy.
In this series, the culture yield of pleural fluid and pleural
children were diagnosed by histopathologic findings of caseat- biopsy culture were not significantly different, but other studies
ing granulomas or visualized acid-fast bacilli in the absence of have found that the latter is more sensitive in terms of culture5
culture confirmation. One child’s isolate was isoniazid monore- and histopathology.6 Gastric aspirates or sputum samples were
sistant; the remaining isolates were susceptible to all first-line culture-positive in 33% of patients who had no evidence of
TB medications. parenchymal involvement by plain radiographs. This indicates
the possible utility of obtaining sputum or gastric aspirate
Treatment
cultures even in children with apparently isolated pleural dis-
Mean duration of therapy was 7.9 months (median: 6 ease, especially in circumstances in which sampling the pleural
months; range: 6 –13 months). Therapy was extended beyond 9
fluid itself is not feasible.
months in 10 children: 5 for nonadherence with initial regimens
There are several limitations to this study. Given the
(before directly observed therapy became available), 3 children
retrospective nature of the study, not all data were available for
with other extrapulmonary disease (2 pericarditis, 1 meningitis), 1
each patient. It was difficult to assess the utility of diagnostic
child with endobronchial disease, and 1 child with acute lympho-
assays such as PCR and adenosine deaminase in this study, as
blastic leukemia. Steroids were administered to 5 patients for a
they were performed in few children. For some children, this
mean of 6.4 weeks. Three children receiving steroids had extrapul-
might have been because TB was not suspected at the time the
monary disease (2 pericarditis, 1 meningitis), 1 had endobronchial
procedure was performed. In other instances, assays were not
TB, and 1 child had extensive bilateral pleural effusions.
available at the time the child was being evaluated. Diagnostic
Prognosis evaluation may have been less aggressively pursued for children
End-of-therapy chest radiographs were abnormal in 28 of 45 with identifiable source cases for whom TB susceptibility infor-
patients (62.2%) (Table 3). The most common abnormalities were mation was known. Finally, for some children, pleural biopsies
residual pleural effusion or thickening (13/28, 46.4%), pulmonary might not have been obtained because they were deemed higher
infiltrate or scarring (8/28, 28.6%), and hilar adenopathy (7/28, risk than thoracentesis. The rationale for diagnostic decision-
25%). All children had complete symptomatic resolution of dis- making was not always entirely evident from the medical record.
ease. The 2 children with TB pericarditis had no evidence of It is unclear why cultures were not attempted in all patients.
constrictive pericarditis on serial echocardiograms during fol- Although cultures obtained from patient specimens were positive
low-up evaluations up to 3 years after the acute episode. The only for 44% of children in this study, cultures of pleural fluid were
child with sequelae was a 10-year-old boy with TB meningitis who attempted for only a minority of patients (9/45, 20%). For some
had hemiparesis and visual loss. Children who had radiographic patients, concomitant parenchymal disease led clinicians to at-
abnormalities at the end of therapy had repeated imaging after 6 to tempt less invasive methods of obtaining cultures, such as sputum
12 months, with approximately 50% having residual abnormalities collection or gastric aspiration. For other children, small pleural
after this time. The most common radiographic anomalies seen effusions and/or identifiable links with a case of infectious tuber-
after 6 to 12 months were pleural thickening (7/14, 50%) and culosis might have led to more conservative diagnostic ap-
pulmonary parenchymal scarring (5/14, 35.7%). No child devel- proaches. In the last decade, more attempts were made to secure a
oped relapse or recrudescence after therapy was discontinued. microbiologic diagnosis by surgical and medical methods. This

© 2009 Lippincott Williams & Wilkins www.pidj.com | 983


Cruz et al The Pediatric Infectious Disease Journal • Volume 28, Number 11, November 2009

was likely driven at least in part by increasing rates of drug- directly observed therapy. The index of suspicion for tuberculosis
resistant tuberculosis. should remain high in children with pleural effusions, with TST
The diagnosis of TB in a child is a sentinel event in a and the presence of epidemiological risk factors facilitating the
community, serving as a marker for recent transmission. This was diagnosis of tuberculosis in the majority of children.
also noted in this series, where 60% of the source cases were
REFERENCES
identified only after the child was diagnosed with disease. How-
ever, the diagnosis of extrapulmonary TB in childhood can be 1. Baumann MH, Nolan R, Petrini M, et al. Pleural tuberculosis in the United
States: incidence and drug resistance. Chest. 2007;131:1125–1132.
challenging. In addition to molecular methods, new surgical tech-
2. Chiu CY, Wu JH, Wong KS. Clinical spectrum of tuberculous pleural
niques might also increase diagnostic yield. VATS has been effusion in children. Pediatr Int. 2007;49:359 –362.
increasingly used in the diagnosis and treatment of children with 3. de Pontual L, Balu L, Ovetchkine P, et al. Tuberculosis in adolescents: a
pyogenic empyemas.7 VATS allows for a minimally invasive French retrospective study of 52 cases. Pediatr Infect Dis J. 2006;25:930 –
exploration of the thoracic cavity. Although few data exist for 932.
utility of VATS in children with pleural TB, potential benefits of 4. Merino JM, Carpintero I, Alvarez T, et al. Tuberculous pleural effusion in
VATS in comparison with conventional thoracotomy include the children. Chest. 1999;115:26 –30.
ability to biopsy the pleura, drain empyemas, and decorticate 5. Valdés L, Alvarez D, San José E, et al. Tuberculous pleurisy: a study of 254
tissue. In one series of 37 patients, no mortality or intraoperative patients. Arch Intern Med. 1998;158:2017–2021.
complications were noted, and full lung re-expansion was achiev- 6. Nance KV, Shermer RW, Askin FB. Diagnostic efficacy of pleural biopsy as
able.8 As VATS becomes used more frequently, its role in the compared with that of pleural fluid examination. Mod Pathol. 1991;4:320 –
324.
diagnosis of children with pleural TB should be evaluated.
7. Schultz KD, Fan LL, Pinsky J, et al. The changing face of pleural empyemas
TB in children is seen concomitantly with parenchymal in children: epidemiology and management. Pediatrics. 2004;113:1735–
disease in a majority of cases. Many children had residual pleural 1740.
scarring or other radiographic abnormalities at the conclusion of 8. Yim AP. The role of video-assisted thoracoscopic surgery in the manage-
therapy, but no relapses were seen with 6-month courses of ment of pulmonary tuberculosis. Chest. 1996;110:829 – 832.

984 | www.pidj.com © 2009 Lippincott Williams & Wilkins

Вам также может понравиться