Pleural Effusion Due to Drug Reactions
Adverse reactions to drugs produce only a small percentage of all pleural effusions. Because the pleural disease in most cases rapidly resolves when the drug is discontinued; however, it is important to consider the possibility of drug-induced pleural disease in all patients with pleural effusions. The lupus-like syndromes associated with various drugs are described in Chapter 21. In this chapter, the pleural diseases resulting from the administration of nitrofurantoin, dantrolene, ergot alkaloids, dasatinib, amiodarone, interleukin 2 (IL-2), procarbazine, methotrexate, interleukin 11 (IL-11), all-trans-retinoic acid, valproic acid and clozapine are discussed. These are the only drugs convincingly incriminated in the production of pleural disease other than drugs that produce the lupus-like syndrome. At the end of the chapter, there is a summary of other drugs that have been implicated but not proved to cause pleural effusions.
NITROFURANTOIN
Nitrofurantoin (Furadantin) is widely used in the treatment of urinary tract infections. Israel and Diamond (1) first reported that the administration of nitrofurantoin could be associated with the development of an acute febrile illness with pulmonary infiltrates and pleural effusion. A subsequent review of the literature and the records of the company that produces nitrofurantoin in 1969 revealed that approximately 200 cases of this syndrome had been reported (2). Now, there have been more than 2,000 cases reported (3). It is thought that nitrofurantoin injures the lung through the production of oxygen radicals (4).
Pulmonary reactions to nitrofurantoin may develop in two distinct patterns characterized by the length of treatment before the development of the syndrome (4). The acute presentation occurs within 1 month of initiating therapy with the drug. The symptoms with the acute presentation include dyspnea, nonproductive cough, and fever. The chest radiograph is usually abnormal. In one series of 335 patients, 186 (56%) had infiltrates, 65 (19%) had infiltrates and effusion, 14 (4%) had only an effusion, and 70 (21%) had a normal chest radiograph (5).
Most patients with acute pleuropulmonary reactions to nitrofurantoin have both peripheral eosinophilia (>350/mm3) and lymphopenia (<1,000/mm3) (6). The only reported pleural fluid analysis showed 17% eosinophils (6).
The chronic syndrome occurs when the patient has been taking nitrofurantoin from 2 months to 5 years and is much less frequent than the acute syndrome. The presentation is insidious, with the gradual onset of dyspnea on exertion and a nonproductive cough (4). Patients with the chronic syndrome always have abnormal chest radiographs; diffuse bibasilar infiltrates are the most common abnormality (5). Pleural effusions, which are less common with the chronic form, occur in fewer than 10% of patients. No patients with the chronic syndrome have had a pleural effusion without an infiltrate (4).
The diagnosis of nitrofurantoin pleuropulmonary reaction should be suspected in all patients with a pleural effusion who are taking nitrofurantoin. If the drug is discontinued, the patient with the acute syndrome usually improves clinically within 1 to 4 days, and the chest radiograph becomes normal within a week (5). Symptoms and signs with the chronic syndrome resolve much more slowly (5).
DANTROLENE
Dantrolene sodium (Dantrium) is a long-acting skeletal muscle relaxant used in treating patients with
spastic neurologic disorders. The chemical structure of dantrolene is similar to that of nitrofurantoin (7,8). The chronic administration of dantrolene can lead to an eosinophilic pleural effusion (7,8). At one of my previous institutions, the Veterans Affairs Medical Center in Long Beach, we saw more than 10 instances of pleural effusions due to dantrolene during a 10-year period. This institution had a large population of patients with spinal cord injury for whom dantrolene was frequently prescribed. In one report, four patients developed an eosinophilic pleural effusion 2 months to 3 years after the initial administration of dantrolene (7). In one case, the eosinophilic effusion developed 12 years after dantrolene therapy was initiated (8). The pleural effusions in all patients were unilateral, and no associated pulmonary infiltrates were seen. In the reported series of four patients, one had a pericardial friction rub and another had a pericardial rub with a pericardial effusion (7). Two of the patients were febrile, and two had pleuritic chest pain.
spastic neurologic disorders. The chemical structure of dantrolene is similar to that of nitrofurantoin (7,8). The chronic administration of dantrolene can lead to an eosinophilic pleural effusion (7,8). At one of my previous institutions, the Veterans Affairs Medical Center in Long Beach, we saw more than 10 instances of pleural effusions due to dantrolene during a 10-year period. This institution had a large population of patients with spinal cord injury for whom dantrolene was frequently prescribed. In one report, four patients developed an eosinophilic pleural effusion 2 months to 3 years after the initial administration of dantrolene (7). In one case, the eosinophilic effusion developed 12 years after dantrolene therapy was initiated (8). The pleural effusions in all patients were unilateral, and no associated pulmonary infiltrates were seen. In the reported series of four patients, one had a pericardial friction rub and another had a pericardial rub with a pericardial effusion (7). Two of the patients were febrile, and two had pleuritic chest pain.
All reported patients have had at least 5% eosinophils in their peripheral blood (7). The pleural fluid is an exudate with normal glucose and amylase levels. The differential white blood cell count on the pleural fluid has revealed at least 35% eosinophils in all cases. When dantrolene is discontinued, the patients improve symptomatically within days, but it takes several months for the pleural effusions to resolve completely. Administration of corticosteroids may accelerate the resolution of the eosinophilic effusion (9). The mechanism by which dantrolene produces the eosinophilic pleural effusion is unknown.
ERGOT ALKALOIDS
Ergot alkaloid drugs such as bromocriptine, ergotamine, dihydroergotamine, nicergoline, pergolide, and dopergine are sometimes used in the long-term treatment of Parkinson’s disease. Ergot derivatives have also been used in the treatment and prophylaxis of migraine and cluster headaches (methysergide, ergotamine). The long-term administration of any of these drugs can lead to pleuropulmonary changes (10,11,12,13,14,15,16,17,18). It has been suggested that the pleural changes are due to increased serotonin levels with a subsequent increase in fibroblast activity (19).
Methysergide (Sansert) is a serotonin antagonist used to treat migraine headaches. The association of methysergide administration with the development of retroperitoneal fibrosis and fibrosing mediastinitis is well established (10). One report described 13 cases of “pleurisy” secondary to methysergide treatment (11). It is not clear what these authors meant by the term pleurisy, but apparently all the patients had pleural effusions or pleural thickening (11). The pleurisy developed 1 month to 3 years after methysergide therapy was initiated, and in five patients, it was bilateral. Only 1 of the 13 patients had concomitant retroperitoneal fibrosis. Although no description of the pleural fluid was included in this report, in another report of a patient with bilateral pleural effusions, the fluid was bloody on one side and clear on the other (12). When methysergide was discontinued, the patients’ symptoms and signs improved. At follow-up 6 months or more after discontinuation of the drug, pleural fibrosis was not detectable or was slight in seven patients, moderate in three patients, and severe in two patients. The two patients with severe fibrosis were those who had continued to take methysergide for the longest period (18 and 36 months) after the onset of this pleurisy (11). Therefore, the occurrence of a pleural effusion or pleural thickening in a patient taking methysergide is a strong indication for the prompt discontinuation of the drug.
Rinne (13) reviewed the chest radiographs of 123 patients taking bromocriptine for Parkinson’s disease and found that 7 patients (6%) had pleural effusions, pleural thickening, and pulmonary infiltrates. As of 1988, there had been a total of 23 patients reported who developed pleuropulmonary disease while taking bromocriptine (14). All the patients were men, and most of them have had a history of long-term cigarette smoking. The prevalence of symptomatic pleuropulmonary disease among individuals taking bromocriptine is 2% to 5% (14). Patients had taken the drug for 6 months to 4 years before symptoms developed. The chest radiograph reveals unilateral or bilateral pleural thickening or effusion with or without pulmonary infiltrates. An occasional patient has only pulmonary infiltrates. Analysis of the pleural fluid reveals an exudate with predominantly lymphocytes and frequently eosinophils (14,19). The erythrocyte sedimentation rate (ESR) is sometimes markedly elevated in these patients (18). There is some suggestion that patients taking bromocriptine-type drugs are more likely to develop pleural disease if they have a history of exposure to asbestos (18,20).
As of 2002, there had been a total of 87 patients reported who developed one or more symptoms of serosal fibrosis after taking pergolide, including 38 cases with pleural effusion and 21 cases with pleural fibrosis (21). The mean age of the patients was 64 years, and approximately three fourths of the patients were men (21). Most patients had been taking the drug for
more than 1 year and had not recovered at the time that they were reported (21).
more than 1 year and had not recovered at the time that they were reported (21).
The natural history of pleuropulmonary disease during treatment with ergot alkaloids is unclear. The disease progresses only in some of the patients who continue taking the drug (14). On discontinuation, most patients improve, but complete resolution of the process is rare. It is recommended that annual chest radiographs be obtained in patients who are taking ergot alkaloids on a long-term basis. If the radiograph reveals pleural or parenchymal infiltrates, strong consideration should be given to stopping the ergot alkaloids and using an alternative drug for the treatment of the Parkinson’s disease.
DASATINIB
Dasatinib is a tyrosine-kinase inhibitor approved for the treatment of BCR-ABL positive chronic myeloid leukemia (CML) and Philadelphia chromosome positive acute lymphoblastic leukemia (ALL) after imatinib failure. The incidence of dasatinib-associated pleural effusions is approximately 20% to 40% (22), which is probably the highest incidence of pleural effusion associated with any drug. Administration of dasatinib once daily is associated with fewer pleural effusions than is administration twice daily and is equally effective (23). In one study (23), 319 patients were randomized to receive dasatinib 140 mg qd or 70 mg bid. The incidence of pleural effusion was 39% in the bid group and 20% in the qd group (23). In a subsequent study (24), 662 patients were randomized to dasatinib 100 mg qd, 50 mg bid, 140 mg qd, or 70 mg bid. In this series (24), the incidence of pleural effusions was lower in the 100 mg qd group (14%) than in the 140 mg qd group (26%), the 70 mg bid group (25%), or the 50 mg bid group (23%). The mean time to the development of the pleural effusion was about 200 days, and it tended to be longer in the groups that received 100 mg per day (24). Pleural effusion was cited as the reason for discontinuing the drug in less than 5% of patients. Peripheral lymphocytosis was more common in the patients that developed pleural effusion than in those that did not (24). Most of the pleural effusions were symptomatic (24).
The characteristics of the pleural fluid secondary to dasatinib therapy have not been well described. Bergeron et al. (25) performed thoracentesis on six patients and reported that the fluid was exudative in all and five were lymphocytic and one was a chylothorax. Quintás-Cardama et al. (26) reported that seven of nine pleural fluids were exudative while two were transudative. Seven of the nine effusions were lymphocyte predominant and one was a chylothorax (26). Pleural biopsy has shown lymphocytic infiltration (25).
The mechanisms responsible for the development of the pleural effusion is unknown. It is thought that it is immune mediated given the predominance of lymphocytes in the pleural effusion, the high incidence of patients who have lymphocytosis, and the lymphocytic infiltrates on the pleural biopsy.