Efficacy of blood cultures in the critically ill surgical patient

Background. Blood cultures are commonly obtained to delineate an infectious process in the ill surgical patient with fever, leukocytosis, or other septic parameters. We studied how often bacteremia was diagnosed, whether a positive blood culture changed therapy, and the cost analysis of this practice. Methods. A heterogenous adult population of 158 patients at high risk for bacteremia was retrospectively reviewed. Blood cultures were not obtained in 37 patients, and thus they were excluded from further study. Results. We obtained 1040 blood cultures in 121 patients. Forty-eight patients (40%) had 122 positive cultures; 20 of these patients had only false-positive cultures. Thus 28 patients (23%) had 82 cultures that represented true bacteremia. Among clinical events, only antibiotic changes and interventions occurred significantly more often as a result of a positive blood culture (p less than or equal to 0.05). No change in therapy occurred in most patients with both positive and negative cultures. Cost for all cultures was $60,058 or $1,251 per positive culture and $1,877 per clinical therapeutic event change. Conclusions. Routine ordering of blood cultures is not cost-effective, rarely alters or provides therapeutic direction, and appears not to affect mortality. Obtaining clinically indicated blood cultures as a secondary rather than a primary diagnostic measure is suggested.