BACKGROUND: Severe sepsis is frequently associated with inflammation, multiple-organ dysfunction syndrome, and the accumulation of excess fluid that can be removed by continuous renal replacement therapy. Appropriate therapy for sepsis requires knowledge of the antibiotic susceptibility of the causative organism. However, bacterial isolation and identification are often unsuccessful. The objective of this study was to determine whether bacteria could be recovered from continuous renal replacement therapy ultrafiltrates of clinically septic patients who were either blood culture positive or repeatedly culture negative. DESIGN: Clinical study. SETTING: Adult medical intensive care unit. SUBJECTS: Six clinically septic, blood culture-positive or repeatedly culture-negative patients and four clinically nonseptic patients. INTERVENTIONS: We performed continuous renal replacement therapy on patients. In addition, we devised an in vitro system mimicking human bacteremia to determine whether bacteria traverse the hemofilter. In these experiments, a reservoir containing reconstituted blood was inoculated with Enterococcus faecalis and hemofiltered. The ultrafiltrates were centrifuged or secondarily filtered for bacterial recovery. MEASUREMENTS AND MAIN RESULTS: Bacterial pathogens were recovered in culture from ultrafiltrates of all the clinically septic patients. Ultrafiltrates of the nonseptic patients were sterile. E. faecalis, the bacterium inoculated into the reconstituted blood, was the only organism recovered from the in vitro-derived ultrafiltrates. CONCLUSIONS: Pathogenic bacteria were shown to traverse the hemofilter. Culture of ultrafiltrates of clinically septic, blood culture-negative patients may be useful in recovery of the etiological organism and confirmation of the clinical diagnosis and management of sepsis.
BACKGROUND: Severe sepsis is frequently associated with inflammation, multiple-organ dysfunction syndrome, and the accumulation of excess fluid that can be removed by continuous renal replacement therapy. Appropriate therapy for sepsis requires knowledge of the antibiotic susceptibility of the causative organism. However, bacterial isolation and identification are often unsuccessful. The objective of this study was to determine whether bacteria could be recovered from continuous renal replacement therapy ultrafiltrates of clinically septic patients who were either blood culture positive or repeatedly culture negative. DESIGN: Clinical study. SETTING: Adult medical intensive care unit. SUBJECTS: Six clinically septic, blood culture-positive or repeatedly culture-negative patients and four clinically nonseptic patients. INTERVENTIONS: We performed continuous renal replacement therapy on patients. In addition, we devised an in vitro system mimicking humanbacteremia to determine whether bacteria traverse the hemofilter. In these experiments, a reservoir containing reconstituted blood was inoculated with Enterococcus faecalis and hemofiltered. The ultrafiltrates were centrifuged or secondarily filtered for bacterial recovery. MEASUREMENTS AND MAIN RESULTS: Bacterial pathogens were recovered in culture from ultrafiltrates of all the clinically septic patients. Ultrafiltrates of the nonseptic patients were sterile. E. faecalis, the bacterium inoculated into the reconstituted blood, was the only organism recovered from the in vitro-derived ultrafiltrates. CONCLUSIONS: Pathogenic bacteria were shown to traverse the hemofilter. Culture of ultrafiltrates of clinically septic, blood culture-negative patients may be useful in recovery of the etiological organism and confirmation of the clinical diagnosis and management of sepsis.