BACKGROUND: Most of malignant pleural effusions are serous but 8-33% of them are bloody. We wanted to evaluate the relationships between gross appearance and pleural CEA level or results of histocytology in malignancy associated pleural effusions. We also tried to reevaluate the meaning of CEA measurement in histocylogically proved or unproved malignancy associated pleural effusions. METHODS: We studied 98 cases of malignancy associated pleural effusions, 50 cases of histocylologically proven malignant effusions and 48 cases of histologically unproven paramalignant effusions. We had observed gross appearance and conventional laboratory values and CEA levees for pleural effusions. RESULTS: 44.9% of malignancy associated effusions were bloody(63.6% of bloody effusions were histstocytologically proven malignant effusion). 65.0% of malignancy associated pleural effusions which have RBCs numbers over 100,000/mm3 were cytologically proven malignant effusions. 72.7% of cytologically proven malignant effusions had increased pleural fluid CEA level over 10 ng/ml. 58.2% of cases with pleural CEA over 10ng/ml had positive results in pleural histocytology. There was no definable relationships between pleural fluid CEA elevation and RBCs numbers and results of pleural fluid cytology. CONCLUSION: About half of the cases with malignancy associated pleural effusions were bloody. Histocytologically proven malignant effusions were more common in bloody effusion than non-bloody effusion (63.6% Vs 38.9%). But increased red blood cell numbers was not associated with positivity of pleural histocytology. Pleural fluid CEA elevation(over 10 ng/ml) was not correlated with positive pleural histocytology. But pleural fluid CEA elevation was rare In nonmalignant pleural effusions, and than pleural CEA measurement in uncertain pleural effusions maybe helpful to distinguishes its origin.