Myocardial dysfunction in patients infected with HIV: prevalence and risk factors.

OBJECTIVES To determine the prevalence of and risk factors for myocardial dysfunction in HIV infection. SUBJECTS 173 patients infected with HIV underwent echocardiography. 119 were current or previous injection drug users, 38 were homosexuals, 10 were haemophiliac patients, and six were heterosexual. MAIN OUTCOME MEASURE Detection of impaired ventricular function. RESULTS 26 patients with abnormalities of ventricular size or function or both were identified. The abnormality was (a) dilated cardiomyopathy in 13 patients (eight homosexuals, three drug users, and two haemophiliacs) with a mean CD4 count of 38 cells/mm3, which accords with end-stage disease (in addition, three patients were identified as having borderline impairment of left ventricular function); (b) left ventricular dilatation without loss of function in a further six patients; and (c) isolated right ventricular dilation in seven patients. Follow up echocardiograms were obtained in 71 patients, 18 of whom had myocardial dysfunction (103 echocardiograms, mean (SD) 2.5 (0.6) scans per patient, mean interval 200 (116) days, range 14-538 days). These showed that in four cases of isolated right ventricular dilatation, one of isolated left ventricular dilatation, and two with borderline left ventricular dysfunction myocardial function subsequently reverted to normal. There was no excess of exposure to zidovudine in the patients with myocardial dysfunction. Similarly, patients with myocardial dysfunction had no serological evidence of excess secondary infection with Toxoplasma gondii and cytomegalovirus. CONCLUSIONS There was a high prevalence and wide range of myocardial dysfunction in HIV positive patients. Dilated cardiomyopathy was a feature of advanced HIV disease and affected all major risk groups for HIV infection. In contrast, isolated dilatation of either ventricle occurred at an earlier stage of HIV infection and, particularly in the case of the right ventricle, often was transient. Neither treatment with zidovudine nor infection with Toxoplasma gondii or cytomegalovirus seemed to be responsible for these findings.