Etiology and Pleural Fluid Characteristics of

Large and Massive Effusions

*
FREE TO
VIEW
Jose Manuel Porcel; Manuel Vives
Author and Funding Information
Chest. 2003;124(3):978-983. doi:10.1378/chest.124.3.978
Text Size: A A A
Article
Tables
References
Abstract
Abstract | Materials and Methods | Results | Discussion | References
Study objective: To report the etiology of large and massive pleural effusions, and to compare
their biochemical fluid characteristics with those of smaller size, and between malignant and
nonmalignant conditions.
Design: Retrospective chart review of all patients undergoing thoracentesis at an academic
medical center in Lleida, Spain, during a 10-year period.
Patients: Posteroanterior chest radiographs were available in 766 patients during the study
period. Large pleural effusions (ie, two thirds or more of the hemithorax without its complete
obliteration) were identified in 70 patients (9%), and massive pleural effusions (ie, hemithorax
was completely opacified) were identified in 93 patients (12%).
Results: A similar etiologic spectrum between large and massive pleural effusions was observed.
The most frequent cause of these pleural effusions was malignancy (89 patients; 55%), followed
by complicated parapneumonic or empyema (36 patients; 22%), and tuberculosis (19 patients;
12%). Compared with nonmalignant pleural effusions, patients with large or massive malignant
pleural effusions were more likely to have pleural fluids with higher RBC counts (18.0
10
9
cells/L vs 2.7 10
9
cells/L, respectively; p < 0.001) and lower adenosine deaminase (ADA)
activity (11.5 vs 31.5 U/L, respectively; p < 0.001), which were the two parameters that were
selected by a stepwise logistic-regression model as independent predictors of malignancy. In
addition, large/massive malignant pleural effusions showed higher median RBC counts (18.0
10
9
cells/L vs 4.3 10
9
cells/L, respectively; p < 0.001), higher lactate dehydrogenase levels
(641 vs 409 U/L, respectively; p = 0.001), lower pH (7.39 vs 7.42, respectively; p = 0.006)
content, but similar cytologic yield (63% vs 53%, respectively; p = 0.171) than smaller malignant
pleural effusions.
Conclusions: The presence of a large or massive pleural effusion enables the clinician to narrow
the differential diagnosis of pleurisy, since most effusions are secondary to malignancy or
infections (either bacterial or mycobacterial). Bloody pleural fluid with low ADA content favors a
malignant condition.
When a pleural effusion is diagnosed in a patient, the need for a timely and systematic
evaluation is indicated. The list of etiologies of pleural effusions is extensive, although
sometimes they can be inferred from the clinical circumstances. Ascertaining the amount of
pleural fluid, as documented by chest radiograph, can be helpful in arriving at a presumptive
cause of the pleural effusion. Thus, there is a general agreement that malignancy is the most
common cause of pleural effusions occupying the entire hemithorax.1

Surprisingly, we have
identified only two series, one in the English language literature2

and the other in the Spanish

language literature,3 that support this notion (from a MEDLINE search conducted among articles
from 1966 to 2002, with the subject heading massive pleural effusion). In addition, the leading
etiology of nonmalignant massive pleural effusions is in dispute.
,
2

3 In the present study, we

intended to evaluate the etiologic spectrum of large and massive pleural effusions in a series
larger than those previously reported, as well as to compare the biochemical characteristics of
pleural fluid among pleural effusions of different sizes. We next attempted to determine what
factors from the pleural fluid analysis were predictive of malignancy in patients with large or
massive pleural effusions.
Materials and Methods
Abstract | Materials and Methods | Results | Discussion | References
We reviewed the medical charts of all patients who had undergone a diagnostic thoracentesis at
the University Hospital Arnau de Vilanova (Lleida, Spain) from June 1992 to June 2002. Clinical,
radiologic, and pleural fluid data were recorded. The size of the effusion was assessed on the
posteroanterior radiograph by visually estimating the area of the hemithorax occupied by pleural
fluid. Pleural effusions were deemed to be nonlarge (ie, slight or of moderate size) if they
occupied less than two thirds of the hemithorax, large if they affected two thirds or more of the
hemithorax without reaching its complete length, and massive if they opacified the entire
hemithorax. Only the measurement of the predominant side was considered in patients with
bilateral effusions. We examined the following pleural fluid analytes: RBC count; leukocyte
count; percentage of neutrophils and lymphocytes; glucose level; protein level; lactate
dehydrogenase (LDH) level; adenosine deaminase (ADA) level; pH; fluid/serum protein ratio;
and fluid/serum LDH ratio.
The causes of pleural effusions were determined by well-established clinical criteria.
Specifically, the criteria for pleural effusions of tuberculous origin were as follows: (1) positive
results for Lwenstein cultures of pleural fluid, sputum, or pleural biopsy specimens; (2) the
presence of a granuloma in a pleural biopsy specimen after excluding other causes of
granulomatous pleuritis; or (3) an exudative lymphocytic effusion with an ADA level of > 40 U/L,
along with a positive tuberculin skin test result and the exclusion of any other potential causes of
pleurisy. A pleural effusion was categorized asmalignant if pleural fluid cytology or pleural biopsy
findings were positive for malignancy (ie, true malignant), or if the patient had a known cancer
with no other explanation for the effusion (ie,paramalignant). The term complicated parapneumonic
effusions (PPEs) referred to those non-purulent-appearing effusions that did not resolve without
chest tube drainage, whereas empyemadescribed the presence of pus within the pleural space.
The terms transudate or exudate were based on the cause of the effusion rather than on the
criteria of Light et al.4 Thus, the category transudatesencompasses those effusions that were
clearly due to congestive heart failure, cirrhosis, or nephrosis. Exudative effusions not
associated with neoplasm, tuberculosis, or pneumonia were classified as other exudates.
Statistical Analysis
Results are reported as medians (quartiles). Comparisons between groups used the
2
and
Fisher exact tests for categoric variables, and the nonparametric Kruskal-Wallis and Mann-
Whitney tests were used for continuous variables. The area under the receiver operating
characteristic curve accurately compared the discriminative properties of pleural fluid analytes.
Demographic and pleural fluid data that distinguished malignant from nonmalignant
large/massive effusions in the bivariate analysis, were entered into a stepwise logistic
regression model in order to identify independent predictors of malignancy. All statistical
comparisons were two-sided and were carried out at the 0.05 significance level. Data were
analyzed with a statistical software package (SPSS, version 10.0; SPSS, Inc; Chicago, IL).
Results
Abstract | Materials and Methods | Results | Discussion | References
Of the 1,051 patients who were evaluated during the study period, 285 were excluded from the
study because either the radiologic data were unavailable or only an anteroposterior radiograph
obtained with the patients in the supine position was performed. In the remaining 766 patients, a
posteroanterior upright chest radiograph was obtained for review. Among this study group, 231
patients had malignant effusions, 150 patients had PPEs (uncomplicated effusion, 52 patients;
complicated pleural effusion or empyema, 98 patients), 113 patients had tuberculous effusions,
114 patients had transudative pleural effusions, and 158 patients had other exudative pleural
effusions.
Seventy patients (9%) exhibited large pleural effusions, and 93 patients (12%) exhibited
massive pleural effusions. Most of these pleural effusions were unilateral (159 of 163 pleural
effusions; 98%). There were 100 men and 63 women, with a median age of 64 years (quartiles,
48 to 75 years). Their underlying diseases are shown in Table 1 . Somewhat more than a half of
large or massive pleural effusions (89 of 163 pleural effusions; 55%) were related to
malignancies. The group of patients with large/massive malignant pleural effusions as a whole
encompassed 58 patients with true malignant effusions and 31 with paramalignant effusions.
The following primary tumors were found: lung, 28 tumors; breast, 19 tumors; unknown, 10
tumors; gynecologic, 7 tumors; hematologic, 7 tumors; GI, 5 tumors; and miscellaneous, 13
tumors (mesothelioma, 4 tumors; head and neck, 4 tumors; kidney, 2 tumors; sarcoma, 1 tumor;
thymoma, 1 tumor; and melanoma, 1 tumor). The second most common cause of large and
massive effusions was PPEs, representing a fifth of the total number of etiologies (36 of 163
patients; 22%) and nearly half of nonmalignant etiologies (36 of 74 patients; 49%). Of note, all
patients in this subgroup had complicated PPEs or empyema, whereas none of the 52
uncomplicated PPEs extended to two thirds or more of the hemithorax. The third most common
cause was tuberculous pleurisy (19 of 163 patients; 12%), and a variety of miscellaneous
causes completed the list, of which hepatic hydrothorax deserves mention.
Overall, 38% of malignant effusions (89 of 231), 37% of complicated PPEs and empyema (36 of
98), 27% of hepatic hydrothoraces (4 of 15), and 17% of tuberculous effusions (19 of 113)
affected two thirds or more of the hemithorax. These percentages changed to 27%, 23%, 7%,
and 8%, respectively, if only a subgroup analysis of massive effusions was considered.
Congestive heart failure (2 of 87 patients; 2%) among others, also may cause this condition.
There were significant differences between the groups of patients with large/massive and
nonlarge effusions regarding the following pleural fluid parameters: RBC count; pH; glucose
level; LDH level; pleural fluid/serum protein ratio; and pleural fluid/serum LDH ratio (Table 2 ).
When the 89 patients having large or massive malignant effusions were compared to the
remaining 142 patients in the malignant effusion population, the former had fluids with higher
median levels of RBCs (18.0 10
9
cells/L vs 4.3 10
9
cells/L, respectively; p < 0.001), higher
medial levels of LDH (641 U/L vs 409 U/L, respectively; p = 0.001), and lower pH (7.39 vs 7.42,
respectively; p = 0.006), yet only the subgroup with massive effusions differed in terms of
glucose fluid concentrations (5.55 mmol/L vs 6.27 mmol/L, respectively; p = 0.006). Similar
results were obtained when only true malignant patients were considered for comparisons.
However, the sensitivity of cytologic examination did not differ between the large/massive and
nonlarge malignant pleural effusion groups (56 of 89 patients [63%] vs 74 of 140 patients [53%];
p = 0.171).
On the other hand, the 36 patients with large or massive PPEs showed higher median leukocyte
counts (13.38 10
9
cells/L vs 2.72 10
9
cells/L, respectively; p = 0.001), higher median LDH
levels (2,438 U/L vs 996 U/L, respectively; p = 0.002), lower pH (6.97 vs 7.34, respectively; p <
0.001), and lower glucose levels (0.28 mmol/L vs 5.00 mmol/L, respectively; p = 0.001) than the
114 patients with PPEs of smaller size. Finally, the leukocyte count was significantly higher in
the pleural fluid of the 94 patients whose tuberculous effusions occupied less than two thirds of
the hemithorax (2.35 10
9
cells/L) than in that of patients with large or massive tuberculous
pleurisy (0.92 10
9
cells/L; p = 0.001).
In comparison to the patients with nonmalignant etiologies, those with malignant pleural
effusions who belonged to the large/massive group were older, and their pleural fluids exhibited
significantly higher values of RBC, glucose, and pH, but lower leukocyte count, LDH level, ADA
level, and pleural fluid/serum LDH ratio (Table 3 ). When these parameters entered a stepwise
logistic regression analysis, a high RBC count and a low ADA level were selected as the most
significant independent predictors of malignancy. Identical results were obtained if only a
subgroup analysis of true malignant effusions had been performed.
Discussion
Abstract | Materials and Methods | Results | Discussion | References
To our knowledge, this is the largest study to address the causes of large and massive pleural
effusions. In 1972, Maher and Berger2

reported on 46 patients with pleural effusions occupying

the entire hemithorax, of whom 31 (67%) had malignant diseases and 15 (33%) had
nonmalignant conditions. Later on, Pedro de Lelis et al3 reviewed 84 cases of massive pleural
effusions, and recognized that most (60 effusions; 71%) were of neoplastic origin. Accordingly,
we also found malignancy to be the leading cause of large and massive effusions in our
population, accounting for 89 of 163 cases (55%). Although similar, this rate (59% if only the
subgroup of patients with massive effusion was considered) appears to be somewhat less
dramatic than those rates mentioned above.
The rate of massive PPEs was much lower in prior publications, in which only four patients
(8.7%)2

and two patients (2.4%)3 were reported. Patient selection might explain this low
incidence. In contrast, PPEs and empyema represented the main nonmalignant etiology (49%)
and the second most common overall etiology (22%) in our series. The absence of patients with
typical or noncomplicated PPEs extending into two thirds or more of the hemithorax supports
the inclusion of large effusions as a risk of poor outcome in patients with PPEs that deserve
drainage of the pleural space.5 Tuberculosis, which was barely encountered (two cases; 4%) as
a cause of massive effusion in the American series,2

appeared as the leading etiology of the

nonmalignant category (13 of 24, 54%) in the previous Spanish study,3 and as the third overall
cause in our population. This is probably due to the still high incidence of tuberculosis in Spain.
In fact, two series from Spain involving 642 and 1,000 consecutive patients, respectively, with
pleural effusions found that tuberculosis was the most frequent etiology (25%)6

and the second

most frequent etiology (15.5%).7

n keeping with the results of Valds et al,8 nearly 20% of our

tuberculous effusions affected two thirds or more of the hemithorax. On the other hand, what we
can infer from the finding of an identical etiologic spectrum between large and massive effusions
is that effusions of both sizes have the same clinical significance.
Massive pleural effusions occur occasionally as a complication of cirrhosis (hepatic
hydrothorax),9congestive heart failure,1 trauma,10

chronic pancreatitis,11

dialysis,12

connective
tissue diseases and vasculitis,13

14 ovarian hyperstimulation syndrome,15

chylothorax,16

or
subarachnoid pleural fistulas,17

among others. We were surprised by the observation of a 74-

year-old woman with a recurrent pulmonary embolism, as demonstrated by helical CT scanning,
that was associated with a massive pleural effusion. She had primary antiphospholipid antibody
syndrome, with pleural malignancy being excluded after numerous laboratory and imaging
studies over a period of 9 months. Of note, in a recent study,18 none of 60 patients with pleural
effusions secondary to pulmonary embolism had effusions that occupied more than one half of
the hemithorax.
Obviously, pleural fluid tests are a useful guideline when assessing the etiology of pleural
effusions, regardless of their size. The differences in pleural fluid biochemical profiles among
nonlarge, large, and massive effusions may in part mirror the relative contributions of diverse
etiologies in each group. Overall, the examination of pleural fluid in large/massive pleural
effusions revealed the presence of biological markers of both inflammation (ie, high LDH levels)
and metabolic activity (ie, low pH and glucose content) compared with smaller pleural effusions,
which may be in the genesis of an increased pleural fluid formation. In addition to pleural
inflammation, increased vascular permeability and leakage play a key role in the development of
exudative pleural effusions. Not surprisingly, malignant and empyema fluids, which represent
the main causes of large/massive effusions, have been consistently found to contain high levels
of vascular endothelial growth factor.19 The determination of whether the absolute
concentrations of this factor and certain cytokines that drive hyperpermeability may affect the
volume of pleural effusion awaits further investigation.
As one might anticipate,2

3 older age and hemorrhagic fluids were associated with malignancy.

The differences in pleural fluid ADA levels between malignant and nonmalignant effusions
probably reflect the high incidence of empyema and tuberculosis, which are two disease entities
that tend to have a high pleural fluid ADA level, in the latter subgroup. Interestingly, effusion size
did not influence the yield of fluid cytology, despite a supposedly greater tumor burden in the
pleural space of malignant massive effusions, as reflected by their lower glucose fluid contents.
One possible criticism of the present study is that 27% of the 1,051 patients who underwent
thoracentesis during the study time period were excluded. However, we do not think that this
selection factor biased our results, because the etiologic distribution of the 766 patients who
composed the studied sample strictly reflects the epidemiology of pleural effusions in our
geographic region.7
,
20
In conclusion, the amount of pleural fluid may be of diagnostic yield in that malignancy is a
strong consideration in the diagnosis of any massive bloody pleural effusion. However, a
significant proportion of benign etiologies, namely, complicated PPEs and empyema, and
tuberculosis should be kept in mind, although the first can be easily distinguished from pleural
metastases by the clinical picture,1 and the second can be distinguished by the ADA fluid
content when the results of cytologic studies are negative.21

22
Abbreviations: ADA = adenosine deaminase; LDH = lactate dehydrogenase; PPE =
parapneumonic effusion
Table 1. Etiology of Nonlarge, Large, and Massive Pleural Effusions in 766 Patients
*

View Large | Save Table

Table 2. Comparison of Demographics and Pleural Fluid Data Among Nonlarge, Large, and
Massive Pleural Effusions
*

View Large | Save Table

Table 3. Comparison of Demographics and Pleural Fluid Data Between Patients With Malignant
and Nonmalignant Effusions of Large/Massive Size
*

View Large | Save Table

References
Abstract | Materials and Methods | Results | Discussion | References
1
Light, RW (2001)Pleural diseases 4th ed. ,33-35 Lippincott Williams & Wilkins. Philadelphia, PA:

2
Maher, GG, Berger, HW Massive pleural effusion: malignant and nonmalignant causes in 46 patients.Am
Rev Respir Dis1972;105,458-460. [PubMed]

3
Pedro de Lelis, F, Ortega, G, Molina, M, et al Derrame pleural masivo: estudio de 84 casosMed Clin
(Barc)1984;82,581-583. [PubMed]

4
Light, RW, Macgregor, I, Luchsinger, PC, et al Pleural effusions: the diagnostic separation of transudates
and exudates.Ann Intern Med1972;77,507-513. [PubMed]

5
Colice, GL, Curtis, A, Deslauriers, J, et al Medical and surgical treatment of parapneumonic effusions: an
evidence-based guideline.Chest2000;118,1158-1171. [PubMed] [CrossRef]

6
Valdes, L, Alvarez, D, Valle, JM, et al The etiology of pleural effusions in an area with high incidence of
tuberculosis.Chest1996;109,158-162. [PubMed]

7
Villena, V, Lopez-Encuentra, A, Echave-Sustaeta, J, et al Estudio prospectivo de 1.000 pacientes
consecutivos con derrame pleural: etiologia del derrame y caracteristicas de los pacientesArch
Bronconeumol2002;38,21-26. [PubMed]

8
Valdes, L, Alvarez, D, San Jose, E, et al Tuberculous pleurisy: a study of 254 patients.Arch Intern
Med1998;158,2017-2021. [PubMed]

9
Degawa, M, Hamasaki, K, Yano, K, et al Refractory hepatic hydrothorax treated with transjugular
intrahepatic portosystemic shunt.J Gastroenterol1999;34,28-31

10
Raney, JL, Carey, MJ Massive hemothorax after low impact blunt chest trauma: a case report.J Ark Med
Soc1998;95,157-159. [PubMed]

11
Muray, S, Porcel, JM, Rubio, M Tratamiento conservador de la Iistula pancreatopleural sin seudoquiste
[letter]Med Clin (Barc)1999;112,48-49

12
Rudnick, MR, Coyle, JF, Beck, LH, et al Acute massive hydrothorax complicating peritoneal dialysis,
report of 2 cases and a review of the literature.Clin Nephrol1979;12,38-44. [PubMed]

13
Mathlouthi, A, Ben Mrad, S, Merai, S, et al Massive pleural effusion in systemic lupus erythematosus:
thoracoscopic and immunohistological findings.Monaldi Arch Chest Dis1998;53,34-36. [PubMed]

14
Miyata, M, Fukaya, E, Takagi, T, et al Two patients with polymyositis or dermatomyositis complicated
with massive pleural effusion.Intern Med1998;37,1058-1063. [PubMed]

15
Tansutthiwong, AA, Srisombut, C, Rojanasakul, A Unilateral massive pleural effusion as the only
principal manifestation of severe ovarian hyperstimulation syndrome.J Assist Reprod Genet2000;17,454-
456. [PubMed]

16
Mokshagundam, SP, Minocha, A Massive chylothorax complicating cirrhosis of the liver.South Med
J1996;89,927-930. [PubMed]

17
DSouza, R, Doshi, A, Bhojraj, S, et al Massive pleural effusion as the presenting feature of a
subarachnoid-pleural fistula.Respiration2002;69,96-99. [PubMed]

18
Romero, S, Hernandez, L, Soler, MJ, et al Biochemical and cytologic characteristics of pleural effusions
secondary to pulmonary embolism.Chest2002;121,465-469. [PubMed]

19
Cheng, DS, Rodriguez, RM, Perkett, EA, et al Vascular endothelial growth factor in pleural
fluid.Chest1999;116,760-765. [PubMed]

20
Hernandez, L, Romero, S Derrame pleural maligno. Porcel, JM eds.Enfermedades de la pleura2002,83-94
Edicions de la Universitat de Lleida. Lleida, Spain:

21
Porcel, JM, Vives, M Adenosine deaminase levels in nontuberculous lymphocytic pleural effusions
[letter].Chest2002;121,1379-1380. [PubMed]

22
Kataria, YP, Khurshid, I Adenosine deaminase in the diagnosis of tuberculous pleural
effusion.Chest2001;120,334-335. [PubMed]