The development of new management techniques for human immunodeficiency virus (HIV) allows for longer latency and survival times. Cardiac complications, including pericardial effusion, are being recognized more frequently [1]. Myocardial involvement in patients with Mycobacterium avium-intracellulare (MAI) is documented [2], but MAI as a cause of pericardial effusion in patients with HIV is described in only a few published cases [3–8]. We present a case that highlights the natural course of an untreated MAI infected pericardial effusion and briefly review the literature on this subject. Case Report. A 29-year-old HIVinfected man (Centers for Disease Control and Prevention stage C3) presented to our hospital with a 3-month history of fever, cough, and dyspnea on exertion. The patient had a long history of noncompliance with his antiretroviral therapy. He had received various combinations of the following medications in the years before admission: zidovudine, didanosine, zalcitabine, and ritonavir. He had not been taking antiretrovirals for an undetermined amount of time at presentation. His CD4 lymphocyte count was 27/mm 3 . The patient’s HIV RNA viral load was elevated. Physical examination revealed resting tachycardia, a third heart sound, and a pericardial friction rub. The chest radiographs demonstrated cardiomegaly. Transthoracic echocardiography revealed a large pericardial effusion with no evidence of tamponade (Fig. 1). A pericardiocentesis yielded 980 mL of hemorrhagic fluid. Studies of the pericardial fluid revealed a white blood count of 4000/cmm, red blood count of 2,279,000/cmm, 87% segs, 8% lymphs, 5% monos, protein of 4.8 g/dL, glucose of 78 mg/dL, lactate dehydrogenase of 161 U/L, and a positive acid-fast bacillus smear. The patient improved and was discharged home. Cultures of the fluid were negative at the time of discharge. Three weeks later, the patient was readmitted with a clinical presentation consistent with a recurrence of his pericardial effusion. Cultures of the initial fluid were reported at this admission and identified MAI sensitive to amikacin, rifabutin, clofazimine, and clarithromycin. Given these results, clarithromycin, ciprofloxacin, and ethambutol were begun. These antimycobacterial drugs were stopped 1 week later after the patient developed dystonic movements. He was treated with a combination of ritonavir and saquinavir at discharge. However, he continued to be noncompliant with his medication regimen. He died 3 months after his initial pericardiocentesis from complications relating to recurrent pericardial disease.FIGURE 1.: Two-dimensional echocardiography demonstrating a large pericardial effusion (arrow).Discussion. The pericardium is the most common site of cardiovascular involvement in HIV infection [9]. Many specific infective causes of pericardial effusion have been reported [10,11], but documented MAI infection is uncommon and is reported in only a limited number of cases [3–8]. Woods and Goldsmith [3] reported a case of a small pericardial effusion (on echocardiography) in a patient with acquired immunodeficiency syndrome (AIDS) whereby MAI was cultured from the pericardium and fluid. Despite antimycobacterial therapy, this patient died of cardiac dysfunction. Cohen et al. [4] described a patient with a moderately sized pericardial effusion on echocardiography. An emergency pericardiocentesis yielded 300 mL of fluid that was culture positive for MAI. The patient expired on the same day as the pericardiocentesis [4]. Another case of MAI pericardial effusion in patients with AIDS, described by Kerns et al. [5], involved a large pericardial effusion detected by echocardiography. Cultures from the pericardial fluid and pericardium were positive for MAI. With surgical drainage and antimycobacterial drugs, including amikacin, ethambutol, rifampin, and ciprofloxacin, the effusion resolved after 2 months [5]. Choo and McCormack [6] described a symptomatic patient with AIDS with a large pericardial effusion noted by echocardiography. Stool, blood, and pericardial fluid cultures confirmed MAI infection. With removal of 1 L of pericardial fluid and treatment with clarithromycin, clofazimine, rifampin, ciprofloxacin, and amikacin, the patient did not develop recurrence [6]. Turco et al. [7] described a case of culture-proven MAI pericardial effusion (approximately 500 cc drained) whereby the patient probably expired from Staphylococcus aureus sepsis soon after presentation. Most recently, Estok and Frances [8] presented a case involving a female with AIDS and a moderate-sized pericardial effusion. Blood, stool, and pericardial cultures were positive for MAI. Despite antimycobacterial treatment with rifabutin, clofazimine, and ethambutol, the patient expired from disseminated MAI and presumed HIV-associated cardiomyopathy. Conclusions. MAI should be considered as an etiological agent in the differential diagnosis of pericardial effusion in patients with AIDS. The clinical findings of pleuritic chest pain, a pericardial friction rub, and cardiomegaly are commonly encountered. Furthermore, the presence of pericardial effusion predicts a poor prognosis in patients with HIV infection, although the size of the effusion is not prognostic. Finally, MAI pericardial effusions may be effectively treated with antimycobacterial therapy, highly active antiretroviral therapy, and surgical management [1]. However, our case suggests a poor prognosis for patients with a pericardial effusion caused by MAI who do not receive antiretroviral and antimycobacterial therapy.