Exudative pleural effusions are common manifestations, but sometimes its causes remain unknown. In Korea, the most common causes of exudative pleural effusion are tuberculosis in 59–64% and malignant neoplasms in 15–31%. However, 10–18% remain unknown^1–3).

In western counties, the most common causes of exudative pleural effusion are malignant neoplasm in 35–68% and tuberculosis in 1–30% (mean 14 %), where 12 – 24 % remain unknown^4–6,8,9).

The identification of the causes of exudate is necessary because treatment and prognosis are determined according to etiologies but some cases remained unknown despite extensive diagnostic studies (chest X-ray, biochemical assay of fluid, bacterial smear and culture, cytology, pleural biopsy and thoracoscopy). In these cases, repeated pleural biopsies and cytology increase the diagnostic yield in patients^6–8) but add to the expense and morbidity of those patients.

The purpose of this investigation was to determine the causes of nonspecific pleuritis on initial pleural biopsies and to determine if individuals with nonspecific pleuritis could be distinguished from those with malignant or tuberculous pleural effusion by routine laboratory data.

Pleural effusion is an important and common clinical finding, but diagnosis of the causes of effusion is often difficult. The diagnosis of pleural effusion is most often established by chemistries of fluid, bacterial smear and culture, cytology for malignancy and closed pleural biopsy.

Diagnostic yield in patients with tuberculous pleurisy had been reported to 20–25% by pleural fluid culture, 60–80% by closed pleural biopsy, 76% by pleural biopsy culture and 95% by combined pleural biopsy with pleural biopsy culture in the same patients^10–12).

The yield of diagnosis of malignant pleurisy was 35–70% based on cytologic finding only and 40–70% on closed pleural biopsy findings only. When these 2 procedures are combined in the same patient, the yield approachs 90%^8,13).

Reasons for the negative results of closed pleural biopsies in those patients who were ultimately diagnosed as having either malignant or tuberculous pleurisy are: 1) early lesion, 2) focal seeding with malignant nodule or granuloma or 3) absence of parietal pleural ivolvement^34,35).

Salyer et al.¹⁸⁾ reports that the causes of exudative pleural effusion in 271 patients were malignancy in 35%, tuberculosis in 20%, and parapneumonic effusion in 16%. Scerbo et al.⁶⁾ notes that out of 163 patients, 40% were malignancy 30% were tuberculosis and 12% were parapneumonic effusion. Prakash et al.⁷⁾ found that the causes of exudate in 414 patients were malignancy in 67.9% and tuberculosis in 1.4%, and Storey et al.⁴⁾ found malignancy in 48% and tuberculosis in 1%.

In summary, the above literature reports that the most common cause of exudate is malignancy in 35–68%, then tuberculosis in 1–30%. Other causes were empyema, connective tissue disease and pulmonary infartion etc. In Korea, the causes of exudative pleural effusion were tuberculosis in 59–64% and variable malignancy in 15–31%^1–3).

Storey et al.^4,6,19) reported that 12–24% of patients with pleural effusion would remain without etiologic diagnosis despite usual diagnostic evaluations. Previous studies have shown that cases of idiopathic pleurisy which were diagnosed to nonspecific pleuritis on initial pleural biopsy, were malignancy in 25–40%. Others were connective tissue disease, pulmonary infarction, tuberculosis, systemic fungal infections, parapneumonic effusion, congestive heart failure, pancreatitis and hepatic disease. However, the indeterminate group remained in the 12–35% range despite extensive studies^14–16). In our investigation, tuberculosis was in 48% and malignancy in 8.7%; others were parapneumonic effusion, subphrenic abscess, congestive heart failure and nephrotic syndrome; 26% remained unknown.

Because the causes of pleural effusion often remain unsolved despite closed pleural biopsy, several biochemical parameters and tumor markers have been measured in pleural fluid to help in the differential diagnosis of various diseases. Liight et al. report that the WBC count and differential WBC count are useful diagnostically^17,18), while Leuallen et al. conclude they are of no use^5,19). Light et al. reported that the presence of predominantly polymorphonuclear leukocytes in pleural fluid indicates that the fluid is the result of acute pleural inflammation, hence raising the possibility of pneumonia with effusion, pulmonary infartion, and pancreatitis, while tuberculous and malignant effusions frequently had more than 50% lymphocytes in the pleural fluid. In this study, the WBC counts of patients with tuberculous pleural effusion were statistically significantly higher than those of patients with malignant pleural effusion in the controls (p<0.005). The study showed a statistically significant (p<0.005) higher value of WBC count in patients ultimately diagnosed as having tuberculous pleurisy than in those ultimately diagnosed as having malignant pleural effusion. Our results are compatible with those of Light et al.¹⁷⁾ in tuberculous pleurisy but not in malignant pleural effution.

In has been a of controversy the glucose concentration in the pleural fluid could aid the diagnosis of the causes of effusion. Caiman et al.²⁰⁾ and Barber et al.²¹⁾ report that a pleural fluid glucose concentration of less than 80 mg/dl suggests a diagnosis of tuberculous pleurisy. Also, Panadero et al.²²⁾ found that in cases with pleural glucose levels less than 60 mg/dl and pleural pH less than 7.30, there was a 90% probability of malignant effusion, while Glenert et al.²³⁾ and Light et al.²⁴⁾ have concluded they are of no use. Our results are in agreement with the latter 2 studies.

Elevated pleural fluid amylase level has been observed in pancreatitis and pancreatic tumor. Light et al.²⁴⁾ in a study of 10 cases of effusion in which amylase levels were more than 160U/dl, report 8 cases were pancreatitis. By Hamm et al’s study²⁵⁾, the mean cholesterol level in malignant effusions was 94 mg/dl, 76 mg/dl in inflammatory effusion and 30 mg/dl in the transudates. The above study indicated that the pleural fluid cholesterol level was an aid in differentiating exudative, especially malignant, from transudative pleural effusions, but our results were not statistically significant.

In 1974, Cortes et al.³⁰⁾ proposed that measurement of the CEA level in pleural fluid suggested malignancy. Thereafter, several authors reported various arbitary levels (CEA levels>12 ng/dl with a diagnostic sensitivity of 34%, Rittger et al.²⁶⁾; >10 ng/dl, 39%, Vladutiu et al.²⁷⁾;20 ng/dl, 65%, Kim²⁹⁾) to separate malignant from benign effusions. But, Mckenna et al.²⁸⁾ found that the CEA level increased in adenocarcinoma only. Pleural fluid adenosine deaminase (ADA) levels in the patients with tuberculous pleurisy were significantly higher than those in the patients with nontuberculous pleurisy, and sensitivity and specificity at various cut-off levels (more than 30 IU/L, Prias et al.³¹⁾; more than 40 IU/L, Lee et al.³²⁾; more than 50 IU/L, Sung et al.³³⁾) were over 90%. In this investign-tion some tests such as amylase, CEA and ADA in the pleural fluid were not performed. Our investigation was limited by the small population of patients studied and by nonperformance of thoracoscopy, pleural biopsy culture and tuberculin skin test.

From our work, we can report the following conclusions: 1) The causes of the nonspecific pleuritis diagnosed on initial pleural biopsy were tuberculosis (11 patients, 48%) and neoplasm (2 patients, 8.7%), while others were parapneumonic effusion, subphrenic abscess, congestive heart failure and nephrotic syndrome; 6 patients (26%) remained unknown. 2) WBC counts (more than 2,000/mm³) and lymphocyte percent (more than 60%) in pleural fluid were significantly elevated in the patients with tuberculous pleurisy compared to malignant pleurisy. 3) In view of the data obtained from our study, we suggest that therapeutic trial with antituberculous medication for patients with a high WBC count (2,000/mm³) and lymphocyte percent (more than 60%) in pleural fluid and who have been diagnosed as having nonspecific pleuritis on initial pleural biopsy can help to locate the etiologies.