Predicting Mycobacterium avium complex bacteremia in patients infected with human immunodeficiency virus: a prospectively validated model
- PMID: 7803630
- DOI: 10.1093/clinids/19.4.668
Predicting Mycobacterium avium complex bacteremia in patients infected with human immunodeficiency virus: a prospectively validated model
Abstract
In cases of advanced infection with human immunodeficiency virus, mycobacterial blood cultures are frequently used to diagnose disseminated infection with the Mycobacterium avium complex (MAC). However, no prospectively validated guidelines exist for the use of such cultures. In this study, a two-part model for predicting MAC bacteremia was developed and then validated prospectively. First, a CD4+ cell count of < or = 50/microL was used to predict bacteremia. Then, among patients with < or = 50 CD4+ cells/microL, the documentation of fever on more than 30 days during the preceding 3 months, a hematocrit of < 30%, or a serum albumin concentration of < 3.0 g/dL was used to predict bacteremia. This model had a sensitivity of 89% and positive and negative predictive values of 30% and 98%, respectively, for the identification of patients with bacteremia. Had the model been applied to patients in this study, the number of blood cultures performed would have decreased by 61%, but 11% of the positive cultures would have been missed. In short, this model can predict MAC bacteremia and can potentially guide the use of mycobacterial blood cultures.
Comment in
-
Yersinia infections: centennial of the discovery of the plague bacillus.Clin Infect Dis. 1994 Oct;19(4):655-61; quiz 662-3. doi: 10.1093/clinids/19.4.655. Clin Infect Dis. 1994. PMID: 7803628 Review. No abstract available.
-
The clinical application of prediction models in medicine--predicting Mycobacterium avium complex bacteremia.Clin Infect Dis. 1994 Oct;19(4):675-6. doi: 10.1093/clinids/19.4.675. Clin Infect Dis. 1994. PMID: 7803631 No abstract available.
Publication types
MeSH terms
LinkOut - more resources
Full Text Sources
Medical
Research Materials
