The febrile parenteral drug user: a prospective study in 121 patients

To determine the most efficient approach to the diagnosis of infective endocarditis (IE) in febrile parenteral drug users (PDUs) and evaluate possible effects of human immunodeficiency virus (HIV) infections or acquired immunodeficiency syndrome (AIDS) on susceptibility to IE and final outcome.

A prospective study of appropriate patients admitted on 149 random sampling days during a 14-month period and review of past experience with IE, HIV, and AIDS admissions to hospital.

An urban university hospital.

Prospectively, 121 febrile PDUs plus an additional 16 found to have IE on nonsampling days during the study period. Retrospectively, all PDUs with IE from 1985 to 1991 and all patients with HIV infections with or without AIDS from July 1990 through December 1991.

Physical examination, hemograms, urinalysis, blood cultures (plus other body fluids when indicated), echocardiography, laboratory testing for HIV status.

Five categories of patients were identified: I. Infective endocarditis (n = 16); II. Other infections with bacteremia (n = 21); III. Bacteremia with unidentified source of infection (n = 14); IV. Infections without bacteremia (n = 52); V. Fever of unknown origin (n = 18). Physical findings and standard laboratory testing did not differentiate Group I from any of the other diagnostic categories. Adding additional IE cases from nonstudy days brought the total to 32. Vegetations were found on echocardiography in 94%; blood cultures, available in 30 of 32 instances, were all positive. HIV or AIDS status was not found to alter susceptibility to IE or influence mortality. While hospital admissions for HIV and especially AIDS have continued to increase among PDUs, the number of cases of IE has decreased since 1988 to 1989.

Based on the high incidence of blood culture positivity and the sensitivity of echocardiography in detecting vegetations in IE, a simple algorithm has been developed for the initial diagnostic management of febrile PDUs admitted with the possible diagnosis of IE. HIV infection, with or without full-blown AIDS, does not appear to affect the incidence or outcome of IE among these patients. Current practices among PDUs may be effecting a decline in IE but not HIV infections.