Rationale for selective use of anaerobic blood cultures.

Because of the declining frequency of anaerobic bacteremia, routinely using half the collected blood volume for anaerobic culture has been challenged. There is no data indicating whether more clinically relevant isolates would be recovered if all or most of the given blood sample were cultured aerobically. In this two-part study, we reviewed cases of anaerobic bacteremia to determine what proportion occurred in situations when anaerobes would be expected and then estimated the yield of different culture approaches by reanalyzing the data from a large prospective clinical blood culture study. The records of 61 patients who had an anaerobic isolate (excluding Propionibacterium species) recovered only from an anaerobic bottle were examined to define clinical settings in which such isolates occur. Fifty-six (92%) patients had clinically important isolates, and the source of infection was obvious at the time of culture in 47 of the 56 (84%). Of 56 patients, 36 (64%) had abdominal signs and symptoms, including 12 with recent abdominal surgery. Of nine patients without an obvious source of infection, six were on high-dose steroids. Relative yields were compared for (i) one aerobic bottle and one anaerobic bottle (5 ml to each) for all blood cultures, (ii) two aerobic bottles (5 ml to each), or (iii) two aerobic bottles plus an extra anaerobic bottle (only for clinically suspected anaerobic sepsis) (5 ml to each). The third approach had the highest yield (475 isolates), because the routine use of two aerobic bottles recovered more Candida spp., members of the family Enterobacteriaceae, and nonfermenters than did the first approach (448 isolates) (P < 0.02), and clinically directed culturing for anaerobes would recover anaerobes missed with the second approach (458 isolates). Our data suggest that the use of two aerobic bottles with selective culturing for anaerobes could increase the number of clinically relevant isolates by at least 6% compared with the current practice of inoculating an aerobic bottle and an anaerobic bottle with equal volumes of blood.