The mechanism by which various agents produce a pleurodesis is unknown. The purpose of this project was to determine whether the pleurodesis that results from the intrapleural administration of talc or doxycycline depends on tumor necrosis factor alpha (TNF alpha). In a randomized, blinded, placebo-controlled study, 34 New Zealand white rabbits were given 400 mg talc or 10 mg/kg doxycycline intrapleurally as a sclerosant through a chest tube. Half the rabbits in each group were also given 2,000 units of ovine, polyclonal affinity purified anti-TNF alpha Fab, or saline as placebo immediately after and 12 h after the injection of the sclerosant. Chest tubes were aspirated at 12-h intervals until their removal at 4 days. Rabbits were killed at 28 days. The pleural fluid volume, cell counts, lactate dehydrogenase (LDH) and pleurodesis scores were compared among groups. Both talc and doxycycline produced an exudative pleural effusion. The pleural fluid volume and the pleural fluid LDH levels were significantly (p < 0.05) greater in the group given doxycycline. The administration of anti-TNF alpha Fab had no significant effect on pleural fluid volume or leukocyte count in either group. However, the administration of anti-TNF alpha Fab resulted in a significant decrease (p = 0.004) in the pleurodesis score for the animals given talc (3.2 +/- 0.8 without Fab and 1.8 +/- 0.9 with Fab). In contrast, the pleurodesis score was virtually identical in the doxycycline group with (3.5 +/- 0.5) and without (3.4 +/- 0.7) Fab. The administration of anti-TNF alpha Fab diminishes the pleurodesis induced by talc but not that resulting from doxycycline. These findings suggest that different mechanisms are involved with the two different sclerosants.