Biomarkers in routine diagnosis of pleural efusions 15
IzvIrnI znanstvenI članek
University Clinic of 
Respiratory and Allergic 
Diseases Golnik, Golnik, 
Slovenia
Korespondenca/
Correspondence:
tiva nemanič, e: nemanic.
tiva@gmail.com
Ključne besede:
biološki označevalci; 
plevralni izliv; lightova 
merila; maligni plevralni 
izliv; paramaligni plevralni 
izliv
Key words:
biomarkers; pleural 
efusion; light’s criteria; 
malignant pleural 
efusion; paramalignant 
efusion
Prispelo: 17. 4. 2017 
Sprejeto: 14. 7. 2017
Biomarkers in routine diagnosis  
of pleural effusions
Uporaba bioloških označevalcev pri rutinskem 
diagnosticiranju plevralnih izlivov
Tiva Nemanič, Aleš Rozman, Katja Adamič, Mateja Marc Malovrh
Abstract
Background: Pleural fluid biochemical analysis is the first step in pleural effusion (PE) diagno-
stics. Our purpose was to analyse the utility of the biomarkers used at our clinic in the routine 
diagnosis of PE.
Methods: We retrospectively reviewed the PE levels of proteins, lactate dehydrogenase (LDH), 
alpha amylase (AA), pH and glucose in 433 patients who were treated at the University Clinic 
Golnik in a one-year period and compared these values with the final identified aetiology of the 
effusions.
Results: The majority of the effusions were determined to be a consequence of malignancy 
(n = 154) or infection (n = 108). In 94 cases the aetiology of the effusions was heart failure and 
in 54 cases other diseases, while 23 effusions remained aetiologically undetermined. Conside-
ring Light’s criteria, the vast majority of the effusions were correctly classified as exudates or 
transudates (97.1 %). Comparing paramalignant and malignant effusions, we detected signifi-
cantly lower values of pleural fluid LDH (p < 0.0005) and proteins (p < 0.0005), and higher pH 
(p < 0.0005) values in the paramalignant effusions.
Conclusion: We have found that pleural LDH and proteins are the most helpful biochemical 
parameters in our routine diagnosis of pleural effusions and helped us to correctly narrow the 
aetiological spectrum. Furthermore, significantly higher pleural LDH and protein values and a 
pH below 7.32 additionally facilitated distinguishing between malignant and paramalignant ef-
fusions. Parameters such as glucose and AA are useful in selected cases and have a limited role 
in routine diagnostics.
Izvleček
Izhodišče: Biokemijska analiza plevralnega izliva (PI) je prvi korak pri diagnosticiranju PI. Na-
men študije je bil ugotoviti uporabnost bioloških označevalcev, ki se uporabljajo pri rutinskem 
diagnosticiranju.
Metode: V retrospektivno analizo smo vključili plevralne vrednosti proteinov, laktat dehidroge-
naze (LDH), alfa amilaze, pH in glukoze 433 bolnikov, ki so bili obravnavani na Univerzitetni klini-
ki za pljučne bolezni in alergijo Golnik v obdobju enega leta, in jih primerjali glede na vzrok izliva.
Rezultati: Vzrok PI so bile v večini primerov maligne (n = 154) ali infekcijske bolezni (n = 108). Pri 
94 bolnikih je bil vzrok srčno popuščanje, v 54 primerih druge bolezni, v 23 primerih pa vzrok 
PI ni bil določen. Z uporabo Lightovih meril je bila večina PI pravilno opredeljena kot transu-
dat ali eksudat (97,1 %). V skupini paramalignih plevralnih izlivov smo zaznali pomembno nižje 
plevralne vrednosti LDH (p < 0.0005) in proteinov (p < 0.0005) ter višje vrednosti ph (p < 0.0005) v 
primerjavi z malignimi plevralnimi izlivi.
Zaključek: Z raziskavo smo potrdili, da z določitvijo plevralnih vrednosti LDH in proteinov po-
membno zmanjšamo spekter diferencialne diagnostike PI, zato sta med najpomembnejšimi bi-
okemičnimi parametri v rutinski diagnostiki. Statistično značilne visoke vrednosti plevralnega 
16 zdrav vestn | januar – februar 2018 | letnik 87
MetaBolne In horMonske Motnje
LDH in proteinov ter vrednosti pH pod 7.32 nam lahko služijo kot dodaten pripomoček pri loče-
vanju malignih od paramalignih izlivov. Določanje glukoze in AA v PI je smiselno in uporabno le 
v nekaterih primerih.
Citirajte kot/Cite as: nemanič t, rozman a, adamič k, Marc Malovrh M. Biomarkers in routine diagnosis of 
pleural efusions. zdrav vestn. 2018;87(1–2):15–21.
DOI: 10.6016/zdravvestn.2579
1. Introduction
Pleural effusions (PEs) are a common 
medical problem and have more than 50 
recognised causes. The most common 
conditions resulting in PEs are conge‑
stive heart failure, pneumonia and ma‑
lignancy (1,2).
Thoracentesis with further specific 
analyses is the cornerstone of diagnosis 
in most cases of pleural effusions, except 
for a clinically and radiologically con‑
vincing heart failure, which responds to 
proper treatment. The first recommen‑
ded step in PE management is determi‑
ning whether the effusion is a transudate 
or an exudate, according to the Light’s 
criteria. If at least one of the following 
criteria is present (ratio of the pleural to 
serum protein values > 0.5, ratio of the 
pleural to serum lactate dehydrogenase 
(LDH) values > 0.6, or pleural LDH valu‑
es > 0.66 upper normal serum level), the 
effusion is classified as an exudate. The 
reported diagnostic accuracy of Light’s 
criteria is 68–95 %, with a sensitivity of 
97.5 % and a specificity of 73.8 % (3‑6). 
Approximately 25 % of transudates are 
erroneously identified as exudates, whi‑
ch occurs most frequently in patients 
with heart failure after receiving diure‑
tics (5).
If effusion is an exudate, additional 
biochemical markers such as pH, gluco‑
se, alpha amylase (AA), cholesterol, and 
triglycerides can be helpful; in selected 
cases, also further cytological and mi‑
crobiological investigations to detect 
carcinoma cells or bacterial strains are 
required (7‑10).
The aim of our study was to evaluate 
the potential utility of each biochemical 
marker used at the tertiary clinic in the 
routine diagnosis of a PE.
2. Method
In this retrospective study, we analysed 
usefulness of biochemical tests in PEs in 
real diagnostic situations and investigated 
the further role of biochemical markers.
We included patients who were tre‑
ated at the University Clinic Golnik, 
in whom a thoracentesis for diagnostic 
purposes was performed in the period 
from 1 Januay 2011 to 31 December 2011. 
Patient’s pleural fluid and serum samples 
were immediately transported to bio‑
chemistry, cytology and microbiology 
laboratories and analysed within one 
hour. Biomarkers in pleural fluid, inclu‑
ding LDH, proteins, pH, glucose and AA, 
were measured using standard routine 
methods. In the cytological laboratory, 
differential blood counts were perfor‑
med, and malignant cells were detected. 
Bacterial cultures, acid‑fast bacillus sme‑
ars and cultures were performed accor‑
ding to culture recommendations.
Three pulmonologists carefully eva‑
luated the diagnostic, treatment and fol‑
low‑up results of the patients in a one‑
‑year period and defined aetiological 
diagnoses of the PEs.
Biomarkers in routine diagnosis of pleural efusions 17
IzvIrnI znanstvenI članek
Pleural malignancy was confirmed 
by the detection of malignant cells in 
the PEs and biopsies obtained through 
a thoracoscopy or a blind pleural bio‑
psy. In polymetastatic lung carcinomas 
with positive computed tomography 
(CT) or positron emission tomography 
(PET) scans of pleural carcinosis, it was 
decided to classify those patients in 
the carcinosis group despite the lack of 
cytological confirmation of carcinoma 
cells. Effusions secondary to lung can‑
cer without any evidence of pleural in‑
vasion were considered paramalignant. 
In acute febrile illness with an ipsilate‑
ral parenchymal infiltrate and resoluti‑
on of the PE with antibiotic treatment, 
the PE was classified as parapneumonic. 
Empyema was diagnosed according to 
either the colonisation of bacteria from 
the PE or a macroscopically purulent 
effusion. Tuberculosis pleuritis was di‑
agnosed if Mycobacterium tuberculosis 
was cultured from the effusion or pleu‑
ral biopsies. A diagnosis of PE secondary 
to pulmonary embolism was determined 
when a pulmonary embolism or infarcti‑
on was observed on CT angiography and 
no other abnormality suggesting pneu‑
monia or cancer was discovered. Heart 
failure (HF) was confirmed by clinical 
and radiological signs of HF (right he‑
art decompensation with elevated cen‑
tral venous pressure, peripheral oedema, 
pulmonary venous congestion, enlarged 
heart, bilateral effusion, a diagnosis of 
previous heart disease and the absence 
of inflammatory pulmonary infiltrates 
or malignancy). Other rarer aetiologies 
were determined according to the pati‑
ents’ history, clinical picture, laboratory 
and imaging tests.
PEs were classified as transudates or 
exudates according to Light’s criteria. 
Furthermore, the values of different bi‑
Table 1: Aetiology of pleural effusions
Aetiology of pleural effusion Number (%)
Malignancy 154 (35.5 %)
Pleural carcinosis 
Mesothelioma
lymphoproliferative diseases 
Paramalignant efusion
84
16
10
44
Infection 108 (25.0 %)
Parapneumonic 
Tuberculosis 
Empyema
77
16
15
Heart failure 94 (21.7 %)
Other 54 (12.5 %)
Pulmonary embolism 
Posttraumatic efusion 
Post-operative efusion
Systemic connective tissue disease 
Trapped lung
Chylothorax 
Ascites
Chronic pancreatitis 
tumour in the right atrium
7
9
11
11
6
4
3
2
1
Idiopathic effusion 23 (5.3 %)
Table 2: Levels of biochemical markers in malignant and paramalignant effusions
Aetiological group /Biochemical 
marker
LDH
[µkat/l]
Proteins [g/l] pH Glucose 
[mmol/l]
Alpha amylase 
[μkat/l]
Malignant effusions 7.5 (± 0.7) 43.0 (± 1.0) 7.30 (± 0.01) 5.83 (± 0.28) 1.39 (± 0.22)
Plural carcinosis 
Mesothelioma
lymphoproliferative disorders
7.8 (± 0.9)
6.4 (± 1.5)
7.3 (± 2.3)
42.8 (± 1.1)
43.1 (± 3.0)
43.9 (± 2.1)
7.31 (± 0.01)
7.21 (± 0.05)
7.35 (± 0.05)
6.26 (± 0.33)
3.44 (± 0.53)
5.50 (± 0.68)
1.65 (± 0.29)
0.58 (± 0.12)
0.61 (± 0.11 = )
Paramalignant effusion 3.3 (± 0.4) 36.0 (± 1.4) 7.39 (± 0.01) 7.44 (± 0.38) 0.68 (± 0.06)
18 zdrav vestn | januar – februar 2018 | letnik 87
MetaBolne In horMonske Motnje
Figure 1a: LDH levels in pleural carciosis, mesothelioma, LPD-
lymphoproliferative disorder and paramalignant effusion. In malignant 
group are all patients with pleural carcinosis, mesothelioma and LPD.
omarkers were compared between the 
different aetiological groups.
In statistical analysis, quantitative 
variables are presented as mean with 
standard error and qualitative variables 
as frequencies and percentages. To as‑
sess the difference in variables between 
the groups we used the Mann‑Whitney 
unpaired t‑test. Correlations were per‑
formed by Pearson’s rank‑order method. 
Statistical analyses were performed with 
Excel and GraphPad Prism 5. We consi‑
dered p‑value of < 0.05 as statistically si‑
gnificant.
3. Result
Out of the 475 patients in whom tho‑
racentesis was performed, 42 patients 
were excluded from the study due to the 
lack of biochemical PE tests. The mean 
age of the 433 remaining patients was 
69.7 (± 0.7) years. In total, 264 patients 
(61 %) were male, and 169 patients (39 %) 
were female.
The most common aetiologies of 
PEs were malignant disease in 154 cases 
(35.5 %), infection in 108 cases (25.3 %) 
and heart failure in 94 cases (21.7 %). 
Other causes were present in 54 cases 
and further division of main three aetio‑
logies are presented in Table 1. In 23 pati‑
ents (5.3 %), the cause of the PE was not 
determined despite extensive evaluation.
LDH and proteins were measured 
in all patients, glucose was measured in 
417 patients (96.3 %), AA in 378 patients 
(87.3 %), pH in 313 patients (72.3 %) and 
haemoglobin in 61 effusions (14.1 %). 
Cytology and microbiology tests were 
performed in all cases.
In accordance with Light’s criteria, 
out of 410 effusions with identified cau‑
ses, 305 (74.4 %) were classified as exu‑
dates, and 105 (25.6 %) as transudates 
with an accuracy of 97.1 %, the positive 
predictive value (PPV) for exudates of 
99.0 % and the PPV for transudates of 
92.1 %. According to the final diagnosis, 
2.9 % of cases were misclassified. In three 
patients with pleural carcinosis, effusi‑
ons were wrongly classified as transuda‑
tes, one of them had pH < 7.30, one had 
normal pH and in one pH was not mea‑
sured. In nine patients with heart failure, 
the effusions were classified as exudates 
according to Light’s criteria. All of these 
patients were receiving diuretic therapy 
prior to thoracentesis, were polymorbid 
and four of these patients had chronic 
renal failure.
Further, pleural LDH values expres‑
sed in µkat/l were significantly eleva‑
ted (7.5 ± 0.7) in malignant compared 
with paramalignant effusions (3.3 ± 0.4) 
(p < 0.0005) (Table 2, Figure 1a). The di‑
fference in pleural protein values expres‑
sed in g/l was also significant, with levels 
of proteins 43.0 ± 1.0 in malignant and 
Biomarkers in routine diagnosis of pleural efusions 19
IzvIrnI znanstvenI članek
Figure 1b: Protein levels in pleural carciosis, mesothelioma, LDH-
lymphoproliferative disorder and paramalignant effusion. Malignant 
group includes all patients with pleural carcinosis, mesothelioma and 
LPD.
36.0 ± 1.4 in paramalignant effusions. 
There were no significant differences in 
LDH or protein levels between different 
types of malignant pleural involvement.
We found significantly lower pH 
values in the PEs in all infectious aeti‑
ologies (6.88–7.29), in the combined 
malignant group (7.30 ± 0.01), pleural 
carcinosis (7.31 ± 0.01) and mesothelio‑
ma group (7.21 ± 0.05). Combined ma‑
lignant effusions had significantly lower 
pH values compared to paramalignant 
PEs (7.39 ± 0.01)(Figure 1c).
Glucose levels, expressed in mmol/l, 
were significantly lower in mesothelio‑
ma (3.4 ± 0.54) and empyema (1.3 ± 0.52) 
compared to other groups. Moreover, glu‑
cose values were lower in all malignant 
effusion compared to paramalignant 
effusion, but the difference was signifi‑
cant only in mesothelioma subgroup. In 
all paramalignant effusions, the glucose 
levels were greater than 3.4 mmol/l.
The AA levels were expressed as µka‑
t/l; considering all pleural malignancies 
combined, we did not find any signifi‑
cant differences between the malignant 
(1.39 ± 0.22) and paramalignant involve‑
ment of the pleura (0.68 ± 0.06).
4. Discussion
The results of this study have con‑
firmed the usefulness of thoracentesis 
and Light’s criteria, which effectively se‑
parate exudative from transudative effu‑
sions. In case of transudate, the pleura is 
unaffected and PE is in most cases a re‑
sult of a HF, nephrosis or liver cirrhosis. 
In contrast, in case of exudate, PE is a re‑
sult of pleural disease, and malignancy is 
suspected, so more extensive diagnostic 
procedure is required (2,3,5,11). Our re‑
sults have shown high accuracy of Light’s 
criteria, which was 97.1 %, being compa‑
rable with previously reported accura‑
cy of 68–95 % (2,6,12). In three patients 
with pleural carcinosis, effusions were 
identified as transudates, which actually 
present only a small proportion of pati‑
ents with pleural carcinosis, but reminds 
us that in highly suspicious cases further 
diagnostics should be performed despite 
biochemical characteristics of a transu‑
date. Additionally, in nine (2.1 %) pati‑
ents with a HF, effusions were misclassi‑
fied as exudates. All of these patients had 
more comorbidities, including chronic 
renal failure, and were using diuretics, a 
situation that was already reported when 
PE in HF presented with characteristics 
of an exudate (5).
In our retrospective one‑year period 
of pleural fluid analyses, we found that 
the most common cause of PE was ma‑
20 zdrav vestn | januar – februar 2018 | letnik 87
MetaBolne In horMonske Motnje
lignancy (32.9 %), followed by infection 
(24.3 %) and HF (23.3 %). Our results 
cannot be directly compared with the 
reported overall proportions of diffe‑
rent aetiologies because this analysis 
excluded a large proportion of patients 
with obvious heart decompensation in 
whom thoracentesis was not performed. 
Moreover, the study was conducted in 
a tertiary centre, where we can expect a 
higher percentage of patients with more 
severe diseases (malignancies and in‑
fections). The resulting proportions of 
different aetiologies among subgroups of 
pleural exudates are comparable to other 
reported studies (2,3,13).
As mentioned previously, pleural bi‑
ochemical test have crucial role in the 
differentiation between exudates and 
transudates, but criteria to separate 
malignant from benign exudates have 
not been established. There are reports 
suggesting that glycosaminoglycans 
(GAGs), VEGF, various tumour markers 
(CA‑125, CEA, CYFRA 21–1 and NSE) 
and mesothelin could be helpful, but the 
accuracy of these markers has not yet 
reached the confidence level for clini‑
cal utility (14‑19). Our study focused on 
biochemical markers, which are already 
routinely used in diagnostics of a PE and 
concentrated on PEs, which are result of 
malignant diseases. We found that the 
levels of proteins and LDH were signifi‑
cantly higher in pleural malignant invol‑
vement compared to paramalignant PEs. 
An additional biochemical marker that 
differed between mentioned groups was 
the level of pH, with significantly lower 
values in effusions with malignant ple‑
ural involvement (combined malignant 
PE, carcinosis and mesothelioma). In 
only one (3.5 %) patient with parama‑
lignant PE pleural pH was lower than 
7.32. Therefore, higher pleural LDH and 
protein levels and lower pH could offer 
additional help in a lung cancer staging 
in borderline cases or where other dia‑
gnostic tools are not feasible. Until now, 
only one study reported difference in 
basic laboratory values between parama‑
lignant and malignant effusions, where 
they detected significantly higher levels 
of serum proteins in patients with ma‑
lignant effusion compared to parama‑
lignant (20).
The clinical importance of other 
biomarkers is limited to specific ca‑
ses. We have not found any significant 
correlations between AA and glucose 
levels and malignant or paramalignant 
effusions.
Figure 1c: pH levels in pleural carciosis, mesothelioma, LDH-
lymphoproliferative disorder and paramalignant effusion. Malignant 
group includes all patients with pleural carcinosis, mesothelioma and 
LPD.
Biomarkers in routine diagnosis of pleural efusions 21
IzvIrnI znanstvenI članek
5. Conclusion
In conclusion, we confirmed the 
accuracy of Light’s criteria for distin‑
guishing transudates from exudates. 
Furthermore, significantly higher pleu‑
ral LDH and protein levels and lower pH 
could be useful in differentiation betwe‑
en paramalignant and malignant PEs.
References
1. Ferreiro l, san josé e, González-Barcala Fj, Álvarez-Dobaño jM, Golpe a, Gude F, et al. Derrames pleu-
rales eosinofílicos: incidencia, etiología y significado pronóstico. archivos de Bronconeumología. 
2011;47(10):504–9.
2. McGrath ee, anderson PB. Diagnosis of Pleural efusion: a systematic approach. american journal of Criti-
cal Care. 2011;20(2):119–28.
3. Bts- British thoracic society. Bts Pleural Disease Guideline 2010. thorax. 2010;65(supp 2):1–76.
4. light rW. Pleural efusions. Medical Clinics of north america. 2011;95(6):1055–70.
5. Porcel jM. Pearls and myths in pleural fluid analysis. respirology. 2010;16(1):44–52.
6. Porcel jM, light rW. Diagnostic approach to Pleural efusion in adults. american Family Physician. 
2006;73(7):1211–1220.
7. Ferrer a, osset j, alegre j, suriñach jM, Crespo e, Fernández de sevilla t, et al. Prospective clinical and 
microbiological study of pleural efusions. eur j Clin Microbiol Infect Dis. 1999;18(4):237–41.
8. joseph j, viney s, Beck B, strange C, sahn sa, Basran Gs. a Prospective study of amylase-rich Pleural efu-
sions With special reference to amylase Isoenzyme analysis. Chest. 1992;102(5):1455–9.
9. sherr hP. origin of Pleural Fluid amylase in esophageal rupture. annals of Internal Medicine. 1972;76(6):985.
10. Good jt, taryle Da, Maulitz rM, kaplan rl, sahn sa. the Diagnostic value of Pleural Fluid ph. Chest. 
1980;78(1):55–9.
11. romero s, Martinez a, hernandez l, Fernandez C, espasa a, Candela a, et al. light’s Criteria revisited: 
Consistency and Comparison with new Proposed alternative Criteria for separating Pleural transudates 
from exudates. respiration. 2000;67(1):18–23.
12. saguil a, Wyrick k, hallgren j. Diagnostic approach to pleural efusion. am Fam Physician. 2014 jul 
15;90(2):99–104.
13. romero-Candeira s, hernández l, romero-Brufao s, orts D, Fernández C, Martín C. Is it Meaningful to use 
Biochemical Parameters to Discriminate Between transudative and exudative Pleural efusions? Chest. 
2002;122(5):1524–9.
14. liang Ql, shi hz, Qin Xj, liang XD, jiang j, Yang hB. Diagnostic accuracy of tumour markers for malignant 
pleural efusion: a meta-analysis. thorax. 2008;63(1):35–41.
15. korczynski P, krenke r, safianowska a, Gorska k, abou Chaz BM, Maskey-Warzechowska M, et al. Diagnostic 
utility of pleural fluid and serum markers in diferentiation between malignant and non-malignant pleural 
efusions. european journal of Medical research. 2009;14(suppl 4):128.
16. Fiorelli a, vicidomini G, Di Domenico M, napolitano F, Messina G, Morgillo F, et al. vascular endothelial 
growth factor in pleural fluid for diferential diagnosis of benign and malignant origin and its clinical appli-
cations. Interactive Cardiovascular and thoracic surgery. 2010;12(3):420–4.
17. alemán C, Manuel Porcel j, Ma segura r, alegre j, esquerda a, ruiz e, et al. Pleural fluid mesothelin for the 
diferential diagnosis of exudative pleural efusions. Medicina Clínica. 2009;133(12):449–53.
18. Duysinx BC, larock M-P, nguyen D, Corhay j-l, Bury t, hustinx r, et al. 18F-FDG Pet imaging in assessing 
exudative pleural efusions. nuclear Medicine Communications. 2006;27(12):971–6.
19. Gu P, huang G, Chen Y, zhu C, Yuan j, sheng s. Diagnostic utility of pleural fluid carcinoembryonic antigen 
and CYFra 21–1 in patients with pleural efusion: a systematic review and meta-analysis. journal of Clinical 
laboratory analysis. 2007;21(6):398–405.
20. Cakmak M, Durkan a. analysis of patients with malignant and paramalignant pleural efusion[cited 2017 
Mar 19]. available from: http://www.alliedacademies.org/articles/analysis-of-patients-with-malignant-
-and- paramalignant-pleural-efusion.pdf