European Journal of Cardiovascular Medicine

The pleural effusion is associated with the history of suffering of mankind. It is a common medical problem encountered by the doctors across the globe. It is a manifestation of respiratory as well as systemic diseases. Tuberculosis and malignancy of lung are predominant causes in India. ^[1]

The pleura is the serous membrane that covers the lung parenchyma, the mediastinum, the diaphragm, and the rib cage. This structure is divided into the visceral pleura and the parietal pleura. The visceral pleura covers the lung parenchyma, not only at its points of contact with the chest wall, diaphragm, and mediastinum but also in the inter-lobar fissures. The parietal pleura lines the inside of the thoracic cavities. ^[2]

Pleural effusion is the abnormal accumulation of fluid in the pleural space⁽²⁾. Pleural fluid accumulates when pleural fluid formation exceeds pleural fluid absorption. Pleural effusion may develop when there is excess pleural fluid formation (from the interstitial spaces of the lung, the parietal pleura, or the peritoneal cavity) or when there is decreased fluid removal by the lymphatics. ^[3]

Pleural effusion occurs in a number of pathological conditions and even exhaustive diagnostic tests fail to reveal the aetiology in as many as 15%–20% of cases. ^[4,5] The classification of pleural effusion as transudative or exudative is the primary diagnostic step because, if the effusion is a transudate, no further diagnostic procedures are necessary, and therapy is directed towards the underlying disease process. However, if the effusion is exudative a more extensive diagnostic work-up is required to distinguish between the many possible causes of exudative effusion. ^[6]

Effusion with normal pleura with intact capillary endothelial barrier are ultrafiltrate of plasma (transudate). Leading causes of transudative pleural effusion are left ventricular failure , pulmonary embolism, cirrhosis of liver. Effusion due to pleural disease leading to increased capillary permeability resemble plasma (exudates). Leading causes are tuberculosis, bacterial pneumonia, cancer, viral infection, pulmonary embolism. When a patient is found to have a pleural effusion, an effort should be made to determine the cause. ^[7]

Tuberculosis is still one of the important causes of exudative pleural effusion in our country. Tuberculosis may affect the pleura at different stages of pulmonary and systemic disease by various mechanisms. ^[8]

In a situation where undiagnosed pleural effusion has come across, the first dilemma needed to be solved is whether the fluid is a transudate or exudates. ^[9] The most frequently used Light‟s criteria, though still considered as the gold standard; sometimes misclassify a transudate as an exudate. However, several prospective studies were unable to reproduce the results obtained by Light et al. In most of these studies, Light’s criteria had a >95% sensitivity for exudates but the specificity was <78%.^[10]

Various causes may be responsible for the presence of cholesterol in the pleural effusion. Few studies have now shown that pleural fluid cholesterol levels differ significantly in transudative and exudative pleural effusion. ^[11] Various studies have shown a high sensitivity and specificity for pleural fluid cholesterol levels to differentiate between transudative and exudative effusion. ^[12]

The present study  is  undertaken  to  evaluate  the   efficacy   of   pleural   fluid   cholesterol  level  and  its  ratio to serum cholesterol level in distinguishing exudative pleural effusion from transudative pleural effusion.

This is a Descriptive, observational single centre study was conducted in the department of General Medicine, Sri Siddartha medical college and hospital, Tumkur, Karnataka over a period one year. Cases of pleural effusion admitted under Medicine department.

Inclusion Criteria:

Patients of either sex above 12 years of age with pleural effusion supported by X ray chest with of tuberculous and nontuberculous origin. Only the results of first thoracocentesis is to be used. Willingness to participate in the study.

Exclusion Criteria:

 Patients meeting the diagnostic criteria of more than one categories, pleural effusions of undetectable or obscure origin, obvious haemothorax secondary to trauma.

A total of fifty patients, under admitted to Department of Medicine having pleural effusion and suffering from varying etiologies, were included in this study. In all cases, a standard clinical protocol was followed, and laboratory tests of pleural fluid were carried out. Only the result of first thoracocentesis was used. Final clinical diagnosis was achieved by standard clinical, radiological and biochemical methods. Invasive procedures like pleural biopsy, lymph node biopsy and CT guided FNAC were done on selected patients. Detection of pleural fluid protein is done by Biuret method, in the department of pathology.

Sample collection: Under all aseptic conditions 10ml of pleural fluid is collected from the patient in a sterile empty vial without any anticoagulant. Plural fluid is made clear by centrifuging at 3000rpm for 10 in in clinical centrifuge.

Protein binds with copper ions in alkaline medium of the Biuret reagent and produces purple colored complex whose absorbance is proportional to the protein concentration. Absorbance of standard, test and blank is measured on the yellow green filter or on a spectrophotometer at 55nm.

Statistical Analysis:

Data were entered in the Microsoft Excel spread sheet and analysed using appropriate statistical technique and software like SPSS. The assumptions of statistical methods need to be satisfactory. The variables included in the analysis of Student's t test have to show a normal distribution. Nonparametric Mann Whitney U test were also done using untransformed values to compare median of exudative and transudative pleural effusion. The cut-off value of PF Cholesterol between exudative and transudative pleural effusion was calculated by ROC curve by plotting (1- specificity) on the X-axis and sensitivity on Y-axis using different cut-off points of PF Cholesterol levels to arrive at the choice which provided the greatest sum of sensitivity and specificity. Using this cut off value sensitivity, specificity, PPV and NPV for PF Cholesterol to differentiate exudative and transudative pleural effusion was calculated and compared with those values of P/S Protein and P/S LDH. Subsequently, ANOVA with Post Hoc Tukey HSD of PF Cholesterol (after square root transformation) and P/S Cholesterol (with original value) in tubercular pleural effusion, empyema, malignant pleural effusion were done to show the significance of mean difference of each other.

The study was carried out on patients of pleural effusion from different medical wards.

Table 1: Distribution of exudative and transudative pleural effusion according to etiology

In the study, among exudative pleural effusion, 10(20%) patients were tubercular,13(26%) had malignancy,4(8%) had empyema,4(8%) had connective tissue disorders,1(2%) had pancreatitis. Among transudative pleural effusion 6(12%) had congestive heart failure,6(12%) had chronic kidney disease,3(6%) had chronic liver disease, 3(%6) had nephrotic syndrome.

Table 2: Pleural fluid glucose level in study population:

In TB, 9 (90%) patients pleural fluid glucose >60mg/dl, 3 (30%) had <60mg/dl. In empyema all cases had glucose<60mg/dl and . In malignancy 12 (93%) patients had PF glucose >60mg/dl and 1 (7%) had <60mg/dl. In CTD all cases had PF glucose >60mg/dl. In acute pancreatitis  1(100%) had PF glucose <60mg/dl and 78% of all transudative pleural effusion  were having PF glucose >60 mg/dl.

Table 3: Pleural fluid cell count study population:

PF cell count >1000/cmm was found in 5(50%) TB patients, 4 (100%) empyema patients, 5(39%) malignancy patients and 4(100%) CTD patient. All transudative pleural effusion, acute pancreatitis, 8 (61%) malignancy patients and 5 (50%) TB patients PF cell count <1000/cmm.

Table 4: Pleural fluid cell count study population

In the study pleural fluid for malignant cells was positive in 6 cases ,culture yielded growth of pyogenic organisms in 4 cases and none of the cases had acid fast bacilli.

Table 5: Pleural fluid Cholesterol level showing non Normal distribution.

Table 6: Mean, SD, p Value of P/S Cholesterol Ratio to Distinguish Exudative Pleural

Effusion & Transudative Pleural Effusion

Mean of P/S Cholesterol of exudative pleural effusion is 0.4456 with SD 0.1450 and that of

transudative pleural effusion is 0.22 with SD 0.10037 and p value is <0.0001.

Table 7: Percentage of Pleural Effusion Misclassified by Light’s Criteria

In the study 5.55% transudates and 9.37% exudates were misclassified by Light’s criteria.

Table 8: sensitivity, specificity, PPV, NPV of P/S protein, P/S LDH,PF Cholesterol, P/S Cholesterol

In the study P/S Protein showed higher sensitivity(92.75%),specificity (88.88%) compared to sensitivity and specificity of P/S Cholesterol 84.37% and  83.33% respectively.PF Cholesterol was least specific(50%),PPV was highest (90%) for P/S Cholesterol. P/S LDH showed 91.87% sensitivity and 83.33% specificity to distinguish between exudative and transudative pleural effusion, but Light;s criteria had higher sensitivity(94.55%),specificity(60.63%),PPV(96.66%) and NPV(85.00%).

Figure 1: ROC Curve for Pleural Fluid Cholesterol to Distinguish Between Exudative and Transudative Pleural Effusion

Figure 2: ROC Curve of P/S Cholesterol  Ratio for Exudative Pleural Effusion & Transudative Pleural Effusion.

Figure shows ROC plot of P/S Cholesterol. Receiver-operating characteristic curve for P/S Cholesterol levels, showing (1 - specificity) on the x-axis and sensitivity on the y-axis using different cutoff points of P/S ALP levels to arrive at the choice of the most appropriate cut-off point and which provided the greatest sum of sensitivity and specificity. Table shows area under curve 0.888 with 95% CI 0.785-0.991.

In the study(Table 1), 32(64%) patients had exudative pleural effusion and 18(32%) patients had transudative pleural effusion. Among exudative pleural effusion, 10(20%) patients were diagnosed to be tubercular, 4(8%) patients had empyema, 13 (26%) patients had malignancy, 4(8%) patients had CTD and 1(2%) patient had acute pancreatitis. Among transudative pleural effusion 6(12%) patients had congestive heart failure, 6(12%) patients had chronic kidney disease, 3 (6%) patients had nephrotic syndrome and 3(6%) patient had cirrhosis of liver. Evidently tubercular pleural effusion  and malignant pleural effusion outnumbered the other etiologies of pleural effusion which is consistent with other Indian studies. ⁽¹³⁾

This study shows increased incidence of empyema in older age group. No malignant effusion was found below 35 years of age, Our study showed that increased incidence of lung cancer below 50 years of age. This is consistent with the fact that median age of diagnosis of lung cancer in India is lower than industrialized nations⁽¹⁴⁾.

In the study, among exudative pleural effusion TB, empyema, malignancy had male predominance, one acute pancreatitis patient was male but all CTD patients were female. Among transudative pleural effusion both CKD and nephrotic syndrome had equal sex distribution. In CHF, male incidence just outnumbered female incidence and all cirrhosis patients were male.

Pleural fluid glucose level in the study (Table 2) showed only 1(10%) patients of tubercular pleural effusion and 1 (7%) patients of malignant pleural effusion had pleural fluid glucose level less than 60mg/dl but 4 (100%) patients of empyema had pleural fluid glucose level less than 60mg/dl and 4(22%) of the transudative pleural effusion had pleural fluid glucose level less than 60mg/dl. The presence of a low pleural fluid glucose is a poor prognostic sign in patients with parapneumonic effusion and serves as an indicator that more aggressive therapy such as tube thoracostomy or thoracoscopy with the breakdown of loculations is necessary⁽¹⁵⁾.

In the present study (Table 3), pleural fluid cell count more than 1000/cmm was found in 5 (50%) tubercular pleural effusion patients, all empyema patients, 5(39%) malignant pleural effusion patients. The only case of acute pancreatitis patient had also less than 1000/cmm cell count in pleural fluid. None of SLE patients and transudative pleural effusion had more than 1000/cmm cell count.

In the present study, non-normal positive skew distribution of PF Cholesterol was normalized after square root transformation (Table 4)⁽¹⁶⁾. Then Student‟s t-test was employed to determine the statistical significance between the means of exudative and transudative pleural effusion. In the group of patients of exudative pleural effusion mean of PF Cholesterol was significantly higher than the mean of transudative pleural effusion (p<0.001).

In this study, distribution of P/S Cholesterol values followed normal distribution pattern so no transformation was done. In the group of patients of exudative pleural effusion the mean of P/S Cholesterol was significantly higher than that of transudative pleural effusion (p<0.001). The optimum cut off value of P/S Cholesterol obtained by ROC curve (Figure 2) was 0.2950 between exudative and transudative pleural effusion with sensitivity 84.37%, specificity 83.33%, PPV 90.00%, NPV 75.00%. Area Under Curve was 0.888 and 95% CI was 0.785-0.991.

There are differences regarding cut-off values of P/S Cholesterol in differentiating exudates and transudates. Hamm et al in their pioneering study in 1987 showed that cut off value of >0.3 would differentiate exudative pleural effusion from transudative pleural effusion with sensitivity 96% and specificity of 94%⁽¹⁷⁾.  Valdes L et al concluded that P/S Cholesterol ratio >0.3 had sensitivity and specificity of 92.7% and 87.6%respectively⁽¹⁸⁾.

ANOVA with Post Hoc Tukey HSD of PF Cholesterol (after square root transformation) there was no significant differences in mean of TB and empyema , TB and malignancy. There was significant difference of mean present between empyema and malignant pleural effusion (0.026). ANOVA with Post Hoc Tukey HSD of P/S Cholesterol with original values showed, significant differences in mean of TB and malignant effusion (0.006),malignant effusion and empyema (0.018). No significant difference were present between TB and empyema.

From this study that Light’s criteria misclassified 1 (5.55%) transudative pleural effusion as exudative pleural effusion and 3(9.37%) exudative pleural effusion as transudative pleural effusion. Comparison of sensitivity, specificity, PPV, NPV of P/S protein ,P/S LDH,PF Cholesterol, P/S Cholesterol showed P/S Protein showed higher sensitivity(92.75),specificity (88.88) compared to sensitivity and specificity of P/S Cholesterol 84.37% and 83.33%. P/S LDH showed 91.87% sensitivity and 83.33% specificity to distinguish between exudative and transudative pleural effusion which was less compared to Light’s criteria.

In our study, based on the statistical analysis it is evident that pleural fluid cholesterol can also be used to distinguish exudative pleural effusion from transudative pleural effusion with cut off level more than 40mg/dl and pleural fluid to serum cholesterol ratio cut off level more than 0.3 suggesting exudative pleural effusion. But it is evident from our study that Light’s criteria has superior sensitivity and specificity compared to pleural fluid cholesterol and its ratio with serum cholesterol to differentiate exudative pleural effusion from transudative pleural effusion.

There is need for large multicentric trials involving large number of cases based on pleural fluid cholesterol or its ratio with serum cholesterol level to differentiate between exudative and transudative pleural effusion in order to come to conclusion.

1. Light, Richard W. Pleural Diseases, Lippincott Williams and Wilkins; 6th Revised edition edition.
2. Fishman's pulmonary diseases and disorders, McGraw-Hill Medical; 5 edition.
3. Harrison’s Principles of Internal Medicine, McGraw-Hill Medical; 19 edition.
4. Sharma SK, Suresh V, Mohan A, Kaur P, Shah P, Kumar A, et al. A prospective study of sensitivity and specificity of adenosine deaminase estimation in the diagnosis of tuberculous pleural effusions. Indian J Chest Dis Allied Sci 2001; 43:149-55.
5. Heffner JE, Sahn SA, Brown LK. Multilevel likelihood ratios for identifying exudative pleural effusion. Chest 2002; 121:1916-20
6. Jimenez Castro D, Diaz Nuevo G, Perez-Rodriguez E. Comparative analysis of Light’s criteria and other biochemical parameters to distinguish exudates from transudates. Rev Clin Esp 2002;202:3– 6
7. Ting ting see,far eastern hospital Tapei Taiwan,Cholesterol in pleural effusions . Chest 1993; 104:646-7
8. Hamm H. Brohan U, Bohmer R, Missmahl HP. Cholesterol in pleural effusions: A diagnostic aid. Chest 1987;92:296–302
9. Valdes L, Pose A, Suarez J, Gonzalez-Juanatey JR, Sarandeses A, San Jose E, et al. Cholesterol: A useful parameter for distinguishing between pleural exudates and transudates. Chest 1991;99:1097–102.
10. Pistolesi M, Miniati M, Giuntini C. Pleural liquid and soluteexchange. Am Rev Respir Dis 1989; 140: 825–47
11. Noppen M, De Waele M, Li R, et al. Volume and cellular content of normal pleural fluid in humans examined by pleural lavage. Am J Respir Crit Care Med 2000; 162: 1023–6
12. Noppen M. Normal volume and cellular contents of pleural fluid. Curr Opin Pulm Med 2001;7:180–182.
13. Albertine KH, Wiener-Kronish JP, Bastacky J, et al. No evidence for mesothelial cell contact across the costal pleural space of sheep. J Appl Physiol 1991;70:123–143.
14. Sahn SA, Willcox ML, Good JT Jr, Potts DE, Filley GF. Characteristics of normal rabbit pleural fluid: physiologic and biochemical implications. Lung 1979; 156: 63–9.
15. Nahid P, Broaddus VC. Liquid and protein exchange. In: Light RW, Lee YC, eds. (Gary): Textbook of pleural diseases. Chapter 3, London: Arnold Publishers, 2003:33-44.
16. Broaddus VC, Araya M, Carlton DP, et al. Developmental changes in pleural liquid protein concentration in sheep. Am Rev Respir Dis 1991;143:38-41 Lai-Fook SJ. Pleural mechanics and fluid exchange. Physiol Rev 2004;84:385-410
17. Wang PM, Lai-Fook SJ. Regional pleural filtration and absorption measured by fluorescent tracers in rabbits. Lung 1999;177:289-309
18. Allen S, Gabel J, Drake R. Left atrial hypertension causes pleural effusion formation in unanesthetized sheep. Am J Physiol 1989;257 2 Pt 2:H690-H692.