[Blood cultures: use of presumptive antiobiograms]

Mortality associated to bacteremia varies between 20 and 40% depending upon several factors, such as focus of infection, microorganism, host conditions, etc. It has also been documented that mortality may double when the patient does not receive antibiotic treatment to which the microorganism is susceptible. The objective of our work has been to determine the correlation between disk diffusion antibiogram according to NCCLS guidelines, from isolated colonies, and the one performed directly from the blood culture flask. During 1996, in the Institute of Cardiology and Cardiovascular Surgery (ICYCC) in Buenos Aires City, 81 episodes of bacteremia were studied. In every case, an antibiogram was carried out: 1) from the bottle: a- Directly (D), harvesting 20 microliters in Mueller Hinton agar, b- Diluted (d), previous centrifugation and Gram staining to adjust turbidity equivalent to 0.5 Mc Farland; 2) from isolated colonies, according to NCCLS guidelines. There were almost no major errors, except with two strains of Enterobacter cloacae versus cephalotin. The diluted method was not so convenient to read inhibition zones, especially with staphylococci. With gram-positive bacteria, the main problems appeared in the direct method with erythromycin, oxacillin and ciprofloxacin because of minor errors. With gram-negative bacteria, major errors were observed in the direct method, mainly with piperacillin (7%) and to a lesser extent with piperacillin tazobactam (2%). Except for imipenem, trimethoprim sulfamethoxazoie and cefotaxime, all antimicrobial agents presented minor errors with both methodologies. Based upon the high rate of minor errors, we consider it is important to confirm results obtained with the standard technique (NCCLS), considering as presumptive those results from the blood culture bottles (D and d).