The yield of bone marrow biopsy and culture compared with blood culture in the evaluation of HIV-infected patients for mycobacterial and fungal infections.

PURPOSE: To compare the clinical utility of bone marrow biopsy and culture specimens with blood cultures for mycobacterial and fungal infections among human immunodeficiency virus (HIV)-infected patients. PATIENTS AND METHODS: All bone marrow biopsies obtained from HIV-infected patients at the University of Alabama at Birmingham (UAB) Medical Center during 1993 to 1995 were blindly reviewed in a standardized format. Bone marrow culture results and blood culture results obtained within 6 weeks of each bone marrow study were compiled. Medical records were reviewed to determine indications for performing bone marrow biopsies, empiric or prophylactic antimicrobial therapies preceding the biopsy, and CD4 counts.
RESULTS: Eighty-two bone marrow studies were obtained from 76 patients. Most were performed during the evaluation of fever, cytopenia, or weight loss. Of 55 bone marrow mycobacterial cultures, 13 yielded Mycobacterium avium complex (MAC) and 2 yielded M tuberculosis (MTB). Of 51 bone marrow fungal cultures performed, 2 yielded Cryptococcus neoformans and 1 Histoplasma capsulatum. All patients with a bone marrow culture positive for MAC had a CD4 count of 20 cells/mm3 or less. The mean CD4 count in this group (+/-95% confidence interval) (8+/-3 cells/mm3) was lower than that of culture-negative cases (41+/-25 cells/mm3); P <0.015). When bone marrow cultures and mycobacterial blood cultures were concurrently obtained, results were usually in agreement between the two sites. The mean time until the report of positive mycobacterial bone marrow cultures (22+/-5 days) was similar to that for blood cultures (24+/-3 days). Most (84%) patients with multiple mycobacterial cultures had completely concordant results (all positive or all negative). When blood or bone marrow culture yielded mycobacteria, only 29% of the corresponding bone marrow examinations revealed stainable acid-fast bacilli (AFB). In contrast, all 3 cases with positive fungal bone marrow cultures also had stainable organisms on histologic examination.
CONCLUSIONS: The combined use of bone marrow biopsy and culture as well as blood cultures provide the maximum diagnostic yield when evaluating patients with AIDS for mycobacterial or fungal infections. However, when mycobacterial infections were diagnosed, bone marrow results seldom provided more immediate or specific information than lysis centrifugation blood cultures. A single lysis centrifugation blood culture should be the first step in the routine evaluation of HIV-infected patients when disseminated MAC infection is suspected.