PEDIATRICS (doi:10.1542/peds.2008-0171)
ARTICLE |
Association Between Parapneumonic Effusion and Pericardial Effusion in a Pediatric Cohort
a Department of Pediatrics
b Department of Radiology, Winthrop-University Hospital, Mineola, New York
c Department of Computer Information Systems and Decision Sciences, St John's University, Queens, New York
d Department of Pediatrics, Albany Medical Center, Albany, New York
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
OBJECTIVE. Associations between pleural and pericardial effusions have been described in malignancy and autoimmune disorders. Bacterial pneumonia is the most frequent cause of parapneumonic effusion; however, knowledge of the relationship between parapneumonic effusion and the presence of pericardial fluid in children is limited. We examined this relationship.
METHODS. We performed a retrospective chart review of pediatric patients who were admitted to our institution during a 6-year period with a diagnosis of either parapneumonic effusion or empyema and who had undergone an echocardiogram, a computed tomography scan of the thorax, or both. All demographic, clinical, radiographic, and laboratory data of these patients were collected, and statistical analysis was done with Student's t tests and 
2 analyses.
RESULTS. We reviewed the charts of 59 children with parapneumonic effusions. Forty-eight underwent 2-dimensional echocardiography, chest computed tomography scan, or both. Of these 48 patients, 54.2% (n = 26) were found to have a concomitant pericardial effusion. The majority of patients with pericardial effusions had left-sided pleural disease. Patients with pericardial effusions had more symptomatic days before hospitalization, lower pleural fluid albumin levels, elevated serum white blood cell counts, elevated pleural fluid white blood cell and absolute neutrophil counts, and an increased incidence of surgical intervention. One patient had evidence of hemodynamic compromise that required pericardiocentesis.
CONCLUSIONS. We found a high incidence of pericardial effusions in pediatric patients with parapneumonic effusions. Leukocytosis, higher pleural fluid leukocyte and neutrophil counts, and a propensity for surgical intervention suggest a prognostic relationship between pericardial effusions and more severe parapneumonic disease. The majority of these pericardial collections resolve with treatment of the underlying pleural disease.
Key Words: pneumonia • pleural effusion • parapneumonic effusion • empyema • pericardial effusion • pericardial thickening • echocardiography • chest CT • chest tube • fibrinolytic therapy • video-assisted thoracoscopic surgery • open thoracotomy
Abbreviations: PCE—pericardial effusion • CT—computed tomography • WBC—white blood cell • VATS—video-assisted thoracoscopic surgery
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In children, pleural effusions are a common complication of bacterial pneumonias.1 Progression of the parenchymal disease into the pleura can lead to the development of complicated parapneumonic effusions and/or empyema if inadequately treated. Parapneumonic effusions are predominantly exudative and occur in up to 60% of patients who are admitted with complicated pneumonia.2
On the contrary, pericardial effusions (PCEs) are reportedly rare in childhood. Some identifiable causes of these effusions include previous cardiac surgery, bacterial pericarditis, malignancy, and connective tissue disorders. In a significant number of cases, it is not possible to identify a cause despite extensive investigations.3
Concurrent pleural effusions and PCEs have been described in connective tissue disorders (eg, lupus, rheumatoid arthritis),4 malignancy (eg, non-Hodgkin's lymphoma),5,6 AIDS,7 congestive heart failure, cardiac surgery, and trauma. In childhood, most pleural effusions are infectious in etiology, with bacterial pneumonia accounting for the majority. With the close proximity of the pleura and pericardium, one may postulate some effect on the pericardial space during pleural infection; however, there are few published data to our knowledge on the relationship between parapneumonic effusions and the presence of pericardial fluid. We describe a retrospective cohort of pediatric patients with a diagnosis of parapneumonic effusion and/or empyema with documented PCE on echocardiography and/or chest computed tomography (CT).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
After approval by the hospital institutional review board, we performed a retrospective chart review of pediatric patients who were aged 0 to 21 years and admitted to our hospital from January 2000 until September 2005. Patients were selected when the chart carried a diagnosis of either parapneumonic effusion or empyema (International Classification of Diseases, Ninth Revision–Clinical Modification codes 510, 510.9, and 511.9). Patients were excluded when they did not receive a cross-sectional imaging study (echocardiography or chest CT) as part of their hospital evaluation. All demographic, clinical, radiographic, and laboratory data and hospital course information were collected and organized by using computer spreadsheets. Statistical data were derived using Student's t test for continuous variables and
2 analysis for categorical measures.
Echocardiograms were reviewed by one pediatric cardiologist and one pediatric echocardiography technician. PCEs were classified as nonexistent, "small" (<100 mL), "moderate" (100–500 mL), or "large" (>500 mL). Chest CT scans were reviewed in a single review session by 3 radiologists to evaluate for the presence and extent of pericardial disease (thickness of the pericardium, pericardial enhancement, and maximum thickness of the pericardial fluid). Effusions 
2 mm were considered small, 2 to 10 mm moderate, and >10 mm large.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
During the study period, a total of 59 pediatric patients were admitted to our institution with a diagnosis of pneumonia with parapneumonic effusion. Forty-eight of the 59 patients underwent echocardiography alone (n = 4), chest CT alone (n = 28), or both (n = 16) and were subsequently analyzed for this study. Of these 48 patients, 26 (54.2%) were found to have a concomitant PCE, 21 small and 5 moderate in size. Of the 26 patients, the presence of PCE was confirmed by both echocardiography and CT for 4 patients, CT alone for 9 patients, and echocardiography alone for 13 patients. Four patients had an echocardiogram without CT scan, and all were found to have PCEs (3 were done to evaluate incidental heart murmurs and the other as an emergent procedure secondary to hemodynamic instability). There were no significant differences between patients with small and moderate effusions in terms of disease severity (based on symptoms, oxygen requirement, and placement in intensive care). All CT scans (n = 44) were performed for the purposes of better evaluating pleural fluid collections and lung parenchyma. Twelve echocardiograms were performed because of PCEs found on CT. Seven patients received an echocardiogram for evaluation of a heart murmur. One patient had an emergency echocardiogram because of acute hemodynamic compromise. None of the echocardiograms demonstrated intracardiac disease. Of the 7 patients who were evaluated for cardiac murmurs, all murmurs were found to be innocent. Figure 1 shows an example of a chest CT scan from 1 of our patients demonstrating both types of effusion. Three patients with PCEs also demonstrated pericardial thickening (average 3 mm), whereas 2 patients without PCE had thickening (average 2.5 mm), which would indicate acute inflammation. There was no difference in the extent of pericardial thickening between patients with left-sided and right-sided disease.
|
During the study period, our hospital averaged 7.8 hospitalizations per year of children with a diagnosis of parapneumonic effusion and/or empyema. Table 1 summarizes the demographic data for our cohort. On average, patients with concomitant pleural effusion and PCE were symptomatic before hospitalization
1.5 days longer than those with only pleural effusions (P < .05). There were no significant differences in age, gender, ethnicity, family history, vaccination use (including influenza and conjugated pneumococcal vaccines), or initial physical examination between the 2 groups. Fifteen patients in the PCE group received antibiotics before hospitalization versus 11 patients in the group without PCE (P value not significant). All patients had community-acquired pneumonia with the exception of 1 patient who had acute lymphoblastic leukemia and developed nosocomial infection with methicillin-resistant Staphylococcus aureus during hospitalization. The majority of the children were previously healthy except for 2 with malignancy, 2 with asthma, and 1 with bronchiectasis. The 2 patients with malignancy developed PCEs.
|
Table 2 summarizes the laboratory data of our cohort. The average pleural fluid white blood cell (WBC) count for patients with PCEs was significantly higher than for patients without (363 x 103 vs 116 x 103, respectively; P = .04). The pleural fluid absolute neutrophil count for patients with PCEs was also significantly higher than for patients without (309 x 103 vs 46 x 103, respectively; P = .03). Pleural fluid albumin was significantly lower in the PCE group (1.54 vs 2.25, respectively; P = .0006) as well. There were no differences between pleural fluid protein, glucose, lactate dehydrogenase levels, or any pleural fluid/serum ratios. In the serum, the average WBC count for patients with PCEs was significantly higher than for patients without (18.1 x 103 vs 11.6 x 103, respectively; P = .01). Serum albumin, protein, and glucose levels did not differ between the groups. Of the 26 patients with PCE, 10 had microbiologic evidence of bacterial infection from the pleural fluid. An organism was isolated in 4 patients (see Table 2), and pneumococcal antigen was isolated in an additional 6 patients. Of the 22 patients without PCE, 6 had evidence of bacterial infection: 4 by culture and 2 additional with pneumococcal antigen.
|
Figure 2 illustrates parapneumonic effusions based on anatomic location. Nineteen (40%) patients had left-sided effusions, 20 (42%) had right-sided, and 9 (19%) had bilateral. Fifty-eight percent (n = 15) of patients in the PCE group demonstrated left-sided pleural effusions, whereas 19% (n = 5) in the group without PCEs had left-sided pleural effusions. Conversely, 18% (n = 4) of patients in the PCE group showed right-sided pleural effusions, compared with 68% (n = 15) in the other group (
2 = 11.3, P < .001). Having bilateral air-space disease did not correlate with the presence of pericardial fluid.
|
Comparison was made between patients who underwent invasive intervention (thoracentesis, chest tube insertion with or without fibrinolytic therapy, video-assisted thoracoscopic surgery [VATS]), or open thoracotomy with or without lobectomy). In the group with PCEs, 81% (n = 21) underwent at least 1 of these mentioned interventions compared with 73% (n = 16) in the other group (P value not significant). Twenty-five percent (n = 7) of patients in the PCE group had an operating room intervention (VATS or open thoracotomy with or without lobectomy, all performed by the same team of 2 pediatric surgeons). Five of these procedures were performed secondary to failed medical management (eg, persistent fever and effusion despite chest tube drainage and antibiotics), whereas 2 were performed as the primary intervention as a result of initial presentation of a complicated pleural effusion (eg, loculated empyema). In the group without PCEs, no patient underwent an operating room procedure (
2 = 6.9; P < .01). Twenty-five (96.1%)of the 26 patients with PCEs were observed without need for pericardial drainage. All 25 patients showed improvement or resolution of their PCEs with treatment of the pneumonia. Twelve patients demonstrated complete resolution via echocardiography before discharge from the hospital with a mean of 2.7 days. The remainder were followed as outpatients and either demonstrated resolution on follow-up echocardiography (n = 5) or were followed clinically and did not have any disease recurrence (n = 8). The patient who required intervention had acute lymphocytic leukemia, sepsis, and multiorgan failure. He underwent drainage of the PCE because of large size and signs of hemodynamic compromise, after which he subsequently improved. Analysis of this patient's pericardial fluid was not available for interpretation.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In our cohort, 54.2% of patients with parapneumonic effusion had a PCE. Previously, Donnelly and Klosterman,8 in a study of 56 pediatric patients who had complicated pneumonia and "noncontributory radiography," noted that 13 (23%) of the 56 patients had PCEs on CT scan. This study did not specifically examine patients with parapneumonic effusions, which may explain the higher incidence in our study. Kishk et al9 investigated the prevalence of cardiovascular abnormalities in 47 children with either acute or chronic empyema by echocardiogram and found that 11 (24%) of the 47 had PCE.
There was an increased occurrence of PCE in children with left-sided parapneumonic effusions. This group also tended to present with air-space disease that involved predominantly the left lower lobe. In 1983, Weiss and Spodick10 reported a cohort of 35 adult patients with pericardial disease and evidence of pleural effusion. Of the pleural fluid collections in their series, 71% were found to be predominantly left-sided.
One possible explanation for the high incidence of pleuropericardial effusions is that of a sympathetic PCE secondary to an adjacent infectious process. Similar pathologic changes were described previously by Kapoor and Shah11 in India with respect to amebic abscess in the liver, with spread to the pericardium. They postulated that the "serous cavity is [likely] involved by the periphery of the inflammatory process surrounding the abscess." In our study, 5 patients who received chest CT scans showed identifiable thickening of the pericardium, which would indicate pericarditis. Donnelly's group8 previously reported that all of their PCEs (13 of 13) were associated with enhancement of the pericardium, suggesting inflammation. The difference may have been attributable to the variable use of intravenous contrast in our study group, which makes it difficult to comment on pericardial enhancement. We theorize that the severity of the pneumonia and the infected pleural fluid may lead to direct inflammation of the pericardium, leading to PCE.
Another possible explanation for the concomitant effusions is the involvement of common lymphatic channels in the left hemithorax draining the left pleural cavity and the pericardial space by the inflammatory process. Parietal pleural lymphatics connect to the pleural space via stomas that are formed by discontinuities in the parietal mesothelial layer.12 The stomas can accommodate particles and have been shown to be the major route of exit of liquid from the pleural space. Riquet et al13 demonstrated that drainage of the pericardial lymphatic vessels was mainly directed toward the tracheobronchial nodes and less frequently toward the prepericardial nodes.
Of patients in our study who underwent pleural drainage, pleural fluid albumin levels were decreased in the group with PCEs compared with the group without (P = .01); however, pleural fluid protein levels, as well as serum albumin and protein levels, showed no such relationship. The reason for these relationships is not clear and may reflect 1 of the weaknesses of this retrospective design.
Although the majority of PCEs in our patients were small (ie, described as "small" on echocardiography, or <10 mm on chest CT), our data suggest that patients with more severe disease are at greater risk for developing concomitant PCE. The patients in the PCE group were symptomatic for a longer time before hospitalization. Serum WBC counts, pleural fluid WBC counts, and pleural fluid neutrophil counts were significantly higher in the group of patients with PCEs. In addition, patients with PCE were more likely to need VATS or open thoracotomy. The majority of these procedures were performed secondary to failed medical management. A recent retrospective review by Padman et al14 revealed that patients who required a CT scan for better evaluation of their disease had a higher likelihood of needing thoracoscopy. Our data may further raise the possibility of using the presence of pericardial fluid as a prognostic factor in patients with parapneumonic effusions. Because almost all of the PCEs improved with appropriate treatment of the underlying pulmonary disease, we believe that appropriate management of the underlying lung disease will lead to resolution of the PCE.
There are limitations to our study. Because this is a retrospective chart review, there are inherent disadvantages in drawing conclusions from the data. First, the differences in the abilities to demonstrate pericardial fluid by the 2 modalities may be attributable to temporal shifting of fluids or because these studies were sometimes performed >24 to 48 hours apart. It was not the aim of this study, however, to compare the 2 techniques but rather to report the findings of pericardial fluid. Second, some CT scans were performed without intravenous contrast, limiting the ability to assess for contrast enhancement and thus assessment of pericardial inflammation. Third, some potentially useful information (eg, exposure to environmental tobacco smoke, conjugated Pneumococcus vaccination history, positive pathogens seen on culture of fluids) was not universally available on every chart and thus could not be adequately evaluated. Fourth, analysis of pericardial fluid was not available for any of our patients. Furthermore, ordered laboratory tests were not consistent from patient to patient, so the power of these analyses was less than would have been ascertained from a prospective design. Last and perhaps most important is the question of whether it is the extent of the pneumonia, the extent of pleural fluid, some combination of the 2, or the overall inflammatory process that is responsible for the pericardial involvement. Our main focus was not to examine the association between pleural fluid size and extent of pneumonia and how this relates to PCEs. To answer this would require a prospective study that examines all cases of pneumonia (with and without pleural involvement) and evaluates the pericardium.
The high incidence of concomitant PCE in children with parapneumonic process described in this study may reflect improved technology for detection and/or a true increase in the incidence of PCE. The frequent occurrence of pericardial involvement should alert the clinician to the importance of close attention to the pericardial structure when interpreting CT imaging of the chest in patients with complex parapneumonic process. Although the majority of small asymptomatic PCEs can be managed conservatively by treating the underlying pulmonary process and these findings do not usually warrant aggressive therapy, this cohort of patients should be observed carefully for signs and symptoms of hemodynamic deterioration. This is especially true for patients with malignancy, immunodeficiency, multisystem organ failure, and sepsis, who are potentially a higher risk group for clinically significant PCEs.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
We found a high incidence of PCEs on echocardiography and/or chest CT in pediatric patients with parapneumonic effusions. The etiology of this observation may relate to a transudative, sympathetic response to the adjacent pleural infection versus a direct inflammatory process that affects lymphatic drainage. Children with left-sided parapneumonic effusions, leukocytosis, higher pleural fluid WBC and neutrophil counts, and low pleural albumin were more likely to develop PCEs. The group who underwent VATS or open thoracotomy all had PCEs. Except for 1 patient, all of the PCEs resolved with treatment of the underlying lung disease.
|
ACKNOWLEDGMENTS |
|---|
This work was performed at Winthrop-University Hospital (Mineola, NY).
|
FOOTNOTES |
|---|
Accepted Aug 4, 2008.
Address correspondence to Jon E. Roberts, MD, 120 Mineola Blvd, Suite 210, Mineola, NY 11501. E-mail: jroberts{at}winthrop.org
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject An association between pleural and pericardial effusions has been described in malignancy and autoimmune disorders, but little is known about this association in the presence of pulmonary infection.
|
| What This Study Adds The detection of pericardial effusions may be used as a prognostic tool for children with parapneumonic effusions and empyema. Pericardial effusions in patients with para-pneumonic effusions and empyema are self-limiting and rarely require intervention.
|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
- Yao CT, Wu JM, Liu CC, Wu MH, Chiang HY, Wang JN. Treatment of complicated parapneumonic pleural effusion with intrapleural streptokinase in children. Chest. 2004;125 (2):566 –571[CrossRef][Web of Science][Medline]
- Strange C, Sahn SA. The definitions and epidemiology of pleural space infection. Semin Respir Infect. 1999;14 (1):3 –8[Medline]
- Mok GC, Menahem S. Large pericardial effusions of inflammatory origin in childhood. Cardiol Young. 2003;13 (2):131 –136[CrossRef][Web of Science][Medline]
- Benseler SM, Silverman ED. Systemic lupus erythematosus. Pediatr Clin North Am. 2005;52 (2):443 –467[CrossRef][Web of Science][Medline]
- Schraader EB, Hesseling PB. Pleuropericardial effusions in children with non-Hodgkin's lymphoma: a report of 2 cases. S Afr Med J. 1987;72 (12):878 –881[Web of Science][Medline]
- Spigland N, Di Lorenzo M, Youssef S, Russo P, Brandt M. Malignant thymoma in children: a 20-year review. J Pediatr Surg. 1990;25 (11):1143 –1146[CrossRef][Web of Science][Medline]
- Mast HL, Haller JO, Schiller MS, Anderson VM. Pericardial effusion and its relationship to cardiac disease in children with acquired immunodeficiency syndrome. Pediatr Radiol. 1992;22 (7):548 –551[CrossRef][Web of Science][Medline]
- Donnelly LF, Klosterman LA. The yield of CT of children who have complicated pneumonia and noncontributory chest radiography.
AJR Am J Roentgenol. 1998;170
(6):1627
–1631
[Abstract/Free Full Text] - Kishk YT, Abou-Elmaqd A, Abdel-Wahab AM. Identification of cardiovascular abnormalities in children with empyema thoracis by two-dimensional and Doppler echocardiography. Chest. 1993;104 (2):405 –410[CrossRef][Web of Science][Medline]
- Weiss JM, Spodick DH. Association of left pleural effusion with pericardial disease. N Engl J Med. 1983;308 (12):696 –697[Web of Science][Medline]
- Kapoor OP, Shah NA. Pericardial amoebiasis following amoebic liver abscess of the left lobe. J Trop Med Hyg. 1972;75 (1):7 –10[Web of Science][Medline]
- Li J. Ultrastructural study on the pleural stomata in human. Funct Dev Morphol. 1993;3 (4):277 –280[Medline]
- Riquet M, Le Pimpec-Barthes F, Hidden G. Lymphatic drainage of the pericardium to the mediastinal lymph nodes. Surg Radiol Anat. 2001;23 (5):317 –319[CrossRef][Web of Science][Medline]
- Padman R, King KA, Iqbal S, Wolfson PJ. Parapneumonic effusion and empyema in children: retrospective review of the duPont experience.
Clin Pediatr (Phila). 2007;46
(6):518
–522
[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2008 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
