By definition, pleural fluid eosinophilia is the presence of an eosinophil count of >10% of total cells in the pleural fluid. Early reports showed that 5% to 8% of exudative pleural effusions are eosinophilic in nature. However, a later study by Rubins and Rubins found that 9.2% of exudative and transudative effusions fulfill the criteria for eosinophilia. Physiology of pleural fluid eosinophilia is not clearly established. In cases of chest trauma and injury, eosinophilia is associated with increased interleukin 5 concentrations in the pleural fluid. In patients with tuberculous pleuritis and chronic eosinophilic pneumonia, granulocyte-macrophage colony-stimulating factor has been shown to be an important stimulator of proliferation and survival of eosinophils in the pleural space.
There are several causes for pleural fluid eosinophilia. Traditionally, the common occurrences have been attributed to introduction of air due to any cause (primary or secondary pneumothorax, thoracentesis, pleural biopsy), or blood (due to infarction or trauma) in the pleural space. Although previous studies had stated that malignancy was an uncommon cause of pleural fluid eosinophilia, the study by Rubins and Rubins invalidated those assertions. They showed that malignancy can be seen in 20.5% of eosinophilic pleural effusions. Other causes of pleural fluid eosinophilia include thoracic surgery, acute pulmonary embolism, benign asbestos pleural effusion, parasitic disease, fungal infection, lymphoma, and drugs. A number of eosinophilic effusions have no identifiable underlying cause.
Numerous drugs can cause a pleural effusion. Examples include procainamide, amiodarone, bleomycin, nitrofurantoin, and interleukin 2. On the other hand, few drugs are associated with blood and pleural eosinophilia. In a review by Morelock and Sahn, only six drugs (propylthiouracil, isotretinoin, bromocriptine, nitrofurantoin, dantrolene, and valproic acid) were named as causing pleural fluid eosinophilia, and only with the latter three was concomitant peripheral eosinophilia also reported.
We found no identifiable cause for pleural and serum eosinophilia in our patient. He was neither taking any of the previously-mentioned medications nor did he report any history of chest trauma. Lupoid drug reaction was dismissed in view of normal ANA titers. However, the patient was taking nifedipine, which in patients with primary pulmonary hypertension has been reported to cause a rapidly reversible pleural effusion and pulmonary edema. Pleural fluid eosinophilia has not been associated with nifedipine. Since the patient was still taking nifedipine at the time of resolution of effusion, we doubt it could have been the cause for pleural fluid eosinophilia.
The temporal relationship between the time of drug intake and onset of eosinophilia may be extremely variable. Ordinarily, it might take hours to days for eosinophilia to appear. Interestingly, in the case of dantrolene sodium, the latency has been as long as 12 years. Resolution of eosinophilia after cessation of the offending agent, which may vary from days to months, usually identifies the underlying cause. In our patient, it took approximately 8 weeks for the pleural effusion to appear and subsequently disappear.
The simultaneous occurrence of serum and pleural fluid eosinophilia and, in particular, their resolution on cessation of fluoxetine may signify a drug-induced hypersensitivity (allergic) reaction. Although nonspecific, one could consider serial measurements of total serum IgE level, which has been reported to increase with an acute allergic drug reaction and to subsequently decrease. We did not measure serum IgE in our patient.
Fluoxetine hydrochloride has been reported to cause lung damage. In humans, it can cause hypersensitivity pneumonitis and pulmonary phospholipidosis. Fluoxetine has also been associated with acute respiratory insufficiency in a patient receiving thyroxine and fluphenazine, possibly through increased pulmonary capillary leakage.[15] To our knowledge, fluoxetine-induced pleural and serum eosinophilia has not been reported. The temporal relation of the onset of pleural fluid and serum eosinophilia with administration of fluoxetine and their resolution on cessation of the drug suggest a drug-induced hypersensitivity (allergic) reaction.