Valproic acid-induced eosinophilic pleural effusion.
* Eosinophilic pleural effusions can result from hemothorax or pneumothorax but can also occur as a consequence of various medications.
* Valproic acid can be associated with the development of eosinophilic pleural effusions.
* Eosinophilic pleural effusions that develop in the setting of valproic acid use should first be treated by discontinuance of the medication before an extensive evaluation.
A 34-year-old white man with a history of schizophrenia and sickle cell trait was admitted for evaluation of fever and nonproductive cough. His recent history included a prolonged hospitalization at an affiliated hospital for fever of unknown origin. He had been admitted to an inpatient psychiatric ward 2 months earlier with an exacerbation of schizophrenia. Treatment included initiation of multiple medications, with a final regimen consisting of valproic acid (1,500 mg/d), clozapine, benztropine, haloperidol, olanzapine, and propranolol.
One month before admission, the patient was transferred to the medical service of another hospital for evaluation of temperatures to 103.2[degrees]F with a declining mental status. At the time of the physical examination, his temperature was 100.4[degrees]F, pulse rate 122 beats/min, respiratory rate 32 breaths/min, and blood pressure 130/75 mm Hg. The patient was lethargic, and physical examination revealed an upper extremity nonblanching macular erythematous rash. Cardiopulmonary, abdominal, and extremity examinations were unremarkable. Chestxray showed bilateral pleural effusions. Laboratory tests revealed a leukocyte count of 12,600/[mm.sup.3] with 28% segmented neutrophils and 40% band forms and a valproic acid level of 113 [micro]g/ml.
The patient was treated with multiple antibiotics (including acyclovir, ceftriaxone, vancomycin, ampicillin, gentamicin, metronidazole, trimethoprim/sulfamethoxazole, azithromycin, and gatifloxacin) for possiblemeningitis and/or pneumonia throughout the hospitalization with little effect on the temperature curve. All antipsychotic agents were withheld throughout the hospitalization. On Day 7 of hospitalization on the medical service, the patient had a transient eosinophilia (cell counts to 3,800/ [mm.sup.3]), which resolved after 1 week.
Lumbar puncture showed four nucleated cells/[mm.sup.3], protein value 36 mg/dl, glucose value 83 mg/dl, negative culture, negative acid-fast stain, and negative Cryptococcus antigen. Results of human immunodeficiency virus enzyme-linked immunosorbent assay and polymerase chain reaction were negative, and stool cultures were negative for Strongyloides species, ova, and parasites.
The patient defervesced (Fig. 1) and was discharged to an inpatient psychiatric ward 13 days later for further management. Discharge medications included haloperidol and lorazepam. During this second inpatient psychiatric stay, he was treated with valproic acid (1,500 mg/ d), haloperidol, and diphenhydramine.
[FIGURE 1 OMITTED]
The patient came to our institution 2 weeks after admission to the psychiatric ward because of recurrent fevers, pleuritic chest pain, and a nonproductive cough. He denied any rash, shortness of breath, headache, myalgias, abdominal pain, or diarrhea.
At physical examination, his temperature was 100.1[degrees]F, pulse rate 122 beats/min, respiratory rate 26 breaths/min, oxygen saturation on room air 94%, and blood pressure 158/76 mm Hg. Cardiopulmonary examination revealed diminished breath sounds bilaterally in the lower half of the lung fields, with associated dullness to percussion, and decreased fremitus. Abdominal and extremity examinations were unremarkable and there was no lymphadenopathy, and no skin lesions were present. Neurologic examination was within normal limits.
Laboratory examination revealed a leukocyte count of 9,900/[mm.sup.3], an eosinophil count of 1,500/[mm.sup.3], and a valproic acid level of 114.7 [micro]g/ml. A chest x-ray revealed bilateral pleural effusions (Fig. 2). Diagnostic thoracentesis yielded 45 ml of an exudative yellow fluid with a leukocyte count of 3,900/[mm.sup.3] (40% eosinophils), a lactate dehydrogenase value of 674 U/L (serum value 221 U/L), protein value 4.8 g/dl (serum value 6.5 g/dl), glucose value 105 mg/dl, and a pH of 7.56. Gram stain revealed no organisms; bacterial and mycobacterial cultures were negative and acid-fast stain was negative. Computed tomography of the chest revealed bilateral pleural effusions without infiltrates, masses, or lymphadenopathy.
[FIGURE 2 OMITTED]
The patient was treated with vancomycin and ceftazidime for possible pneumonia, and the valproic acid was discontinued. The patient defervesced, and the eosinophilia resolved (Fig. 1). Laboratory tests were negative for serum Coccidioides and Toxocara antibodies and moderately positive for Aspergillus. Immunoglobulin E (IgE) was at 2.9 kU/L. Morning cortisol level was 26.3 [micro]g/dl. A PPD tuberculin skin test revealed no in-duration with a reactive anergy panel.
The patient's condition continued to improve after completion of antibiotic therapy, and he was transferred to inpatient psychiatric care after 13 days. Treatment with valproic acid has not been restarted, and the patient has not had recurrent fevers. A follow-up chest x-ray obtained 1 month after admission to our hospital showed significant improvement (Fig. 3).
[FIGURE 3 OMITTED]
Eosinophilic pleural effusions, defined as effusions with eosinophil counts greater than 10% of the total fluid nucleated cell count, account for 5 to 8% of exudative pleural effusions. (1) The presence of eosinophils confers a more benign prognosis in comparison with other exudative effusions, with average survival more than twice as long. (1) Pleural fluid eosinophilia most commonly arises from hemothorax or pneumothorax (2) but can also result from asbestosis, (3) pulmonary infarction, (4) parapneumonic effusion, (1,4) tuberculosis, (4) parasitic infections including toxocariasis, (5) and strongyloidiasis, (6,7) eosinophilic pneumonia, (8) malignancy, (1) and medications. (2) Valproic acid has been reported to be the cause of pleural fluid eosinophilia in two previous cases. (2,9) We report another case of pleural fluid eosinophilia attributable to the use of valproic acid.
Although pleural fluid eosinophilia can result from blood or air in the pleural space, other possibilities, including reactions to medications, must be considered, as 25% of eosinophilic pleural effusions are idiopathic. (1) Medications associated with pleural fluid eosinophilia include dantrolene, (10) mesalamine, (11) fluoxetine, (12) gliclazide, (13) and isotretinoin. (14,15) Our search of the English-language literature yielded only one case report of valproic acid leading to the development of eosinophitic pleural effusion. (2)
Our patient's recurrent bilateral eosinophilic pleural effusions, accompanied by systemic eosinophilia and fever, were associated with the use of valproic acid. Our case is unique because there was an inadvertent rechallenge with valproic acid and the patient's symptoms recurred, therefore lending more evidence to the association. Furthermore, an extensive evaluation did not reveal any of the other known causes of eosinophilic pleural effusion. Specifically, there was no evidence of pneumothorax, hemothorax, pulmonary infiltrates, lymphadenopathy, or mycobacterial, fungal, or parasitic infection. Finally, the two medications common to both hospitalizations were valproic acid and haloperidol; the latter has not been associated with the development of pleural effusions. In addition, haloperidol was given to our patient at a later date, without the development of recurrent symptoms.
Valproic acid is a carboxylic acid used in the treatment of seizure disorders and has more recently been used as a mood stabilizer for selected patients with psychiatric disease. As the use of this agent increases, physicians should increase their awareness that it may cause eosinophilic pleural effusion. If a patient taking valproic acid presents with eosinophilic pleural effusion, we recommend prompt discontinuance of the drug and careful observation. If improvement is noted, the patient can forego an extensive evaluation for other causes and should not be rechallenged with this medication.
(1.) Rubins JB, Rubin, HB. Etiology and prognostic significance of eosinophilic pleural effusions: A prospective study. Chest 1996;110:1271-1274.
(2.) Kaufman J, O'Shaughnessy IM. Eosinophilic pleural effusion associated with valproic acid administration. South Med J 1995;88:881-882.
(3.) Mattson SB. Monosymptomatic exudative pleurisy in persons exposed to asbestos dust. Scand J Respir Dis 1975;56:263-272.
(4.) Martinez-Garcia MA, Cases-Viedma E, Cordero-Rodriguez PJ, Hidalgo-Ramirez M, Perpina-Tordera M, Sanchis-Moret F, et al. Diagnostic utility of eosinophils in the pleural fluid. Eur Respir J 2000;15:166-169.
(5.) Jeanfaivre T, Cimon B, Tolstuchow N, De Gentile L, Chabasse D, Tuchais E. Pleural effusion and toxocariasis. Thorax 1996;51:106-107.
(6.) Goyal SB. Intestinal strongyloidiasis manifesting as eosinophilic pleural effusion. South Med J 1998;91:768-769.
(7.) Emad A. Exudative eosinophilic pleural effusion due to Strongyloides stercoralis in a diabetic man. South Med J 1999;92:58-60.
(8.) Samman YS, Wali SO, Abdelaal MA, Gangi MT, Krayam AB. Chronic eosinophilic pneumonia presenting with recurrent massive bilateral pleural effusion: Case report. Chest 2001;119:968-970.
(9.) Tonnelier JM, Mocquard Y, Broch C, Vandhuick O, Goas P, Goas JY. Febrile pleuropericarditis during valproic acid treatment [in French]. Presse Med 2001;30:62-63.
(10.) Felz MW, Haviland-Foley DJ. Eosinophilic pleural effusion due to dantrolene: Resolution with steroid therapy. South Med J 2001;94:502-504.
(11.) Trisolini R, Dore R, Biagi F, Luinetti O, Pochetti P, Carrabino N, et al. Eosinophilic pleural effusion due to mesalamine: Report of a rare occurrence. Sarcoidosis Vasc Diffuse Lung Dis 2000;17:288-291.
(12.) Behnia M. Dowdeswell I, Vakili S. Pleural fluid and serum eosinophilia: Association with fluoxetine hydrochloride. South Med J 2000;93:611-613.
(13.) Tzanakis N, Bouros D, Siafakas N. Eosinophilic pleural effusion due to gliclazide. Respir Med 2000;94:94.
(14.) Milleron BJ, Valcke J, Akoun GM, Mayaud CM. Isotretinoin-related eosinophilic pleural effusion. Chest 1996; 110:1128.
(15.) Bunker CB, Sheron N, Maurice PD, Kocjan G, Johnson NM, Dowd PM. Isotretinoin and eosinophilic pleural effusion. Lancet 1989;1:435-436.
From the Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, and the VA Connecticut Health Care System, West Haven, CT.
Reprint requests to Jeffrey D. Kravetz, MD, 278 W. Elm Street, New Haven, CT 06515. Email: firstname.lastname@example.org
Accepted February 28, 2002.