Routine Anaerobic Blood Cultures: Back Where We Started?

The debate about whether and when anaerobic blood cultures should be routinely performed in the hospital setting has been ongoing for >15 years [1–3]. Previously, reports that the incidence of anaerobic bacteremia was decreasing, that most positive culture results correlated with predictable clinical findings, and that routine use of anaerobic cultures rarely resulted in clinically relevant diagnostic or therapeutic gains compelled some authors to argue that only selective ordering of these tests should be considered [1, 3, 4]. Selective cultures were proposed only for patients considered to be at high risk for anaerobic infection, as determined by the clinician. This would require the development of guidelines or policies to properly select at-risk patients, as well as some determination of the feasibility of this approach. To date, despite calls to selectively obtain anaerobic blood cultures and to consider additional studies on the subject, these recommendations have not been adopted in adult patients. Although still debated, routine use of anaerobic blood cultures in pediatric patients is not widely practiced and is generally justified by the lower incidence and risk factors for anaerobic infection among children and the sparing of needless phlebotomy [5, 6]. The wisdom of inaction on this debate may have been well served, as evidenced by the article in this issue of Clinical Infectious Diseases by Lassmann et al. [7], which reports the reemergence of anaerobic bacteremia at the Mayo Clinic (Rochester, MN). In their study, which compares 2 periods (1993–1996 vs. 2001–2004), there was a 74% overall increase in the incidence of positive blood culture results for obligate anaerobes per 100,000 patient-days, which was highly significant (P < .001). In addition, there was a 30% increase in the number of positive anaerobic blood culture results per 1000 cultures performed during the 12-year period (P = .002), with a proportional decrease in positive culture results for aerobic and facultative anaerobic organisms. This is in contrast to the 45% decrease in the incidence of anaerobic bacteremia during 1974–1988, which was reported from the same institution in 1991 and spurred much of the original debate [4]. What is the cause for this increasing trend? The authors considered several factors, including a change in culture media and type, a potential shift in the distribution of organisms, and underlying clinical conditions. Although none of those factors reached statistical significance, trends were identified among the relatively small number of charts available for review. Regardless, there are several reasons why laboratories should employ routine anaerobic blood cultures. The first reason to consider is the anaerobic pathogen and the potential for antibiotic resistance. In their study, Lassmann et al. [7] identified Bacteroides fragilis group members and Clostridium species as the most numerous anaerobes isolated. This is consistent with their previous reports, as well as those of others [4, 8–11]. During the 12-year period of their study, both the B. fragilis group and Clostridium species constituted 42%–80% of all anaerobic isolates, depending on the year of isolation. Although the authors did not report the species within these groups, their identification can be potentially important for the selection of empirical antibiotic agents. This is particularly true for members of the B. fragilis group, as evidenced by the most recent report from the national survey of resistance of these organisms, which shows that only 50% of 580 isolates were B. fragilis, and 36% were either Bacteroides distasonis, Bacteroides thetaiotaomicron, or Bacteroides vulgatus [12]. These 3 species are the other most frequently isolated members of the B. fragilis group and are generally the most resistant to multiple antibiotic agents [12–14]. This raises the second reason to continue anaerobic blood cultures in the form of a question: is antibiotic resistance correlated with worse clinical outcomes for anaerobic bacteremia? Indeed, this question has been answered, at least for the B. fragilis group. Several retrospective reviews of infection with Bacteroides species [16–19] and the landmark article by Nguyen et al. [15] in 2000 clearly support this position. In the article by Nguyen and colleagues, a prospective observational study of 128 cases of bacteremia involving the B. fragilis group, mortality, microbial persistence, and clinical failure occurred more frequently among patients who did not receive an appropriate antibiotic agent to treat infection with resistant members of the B. fragilis group [15]. Although isolation and identification are neither practical nor readily available for the individual patient, knowledge of the susceptibility of anaerobic antibiotics at individual hospitals may be important in the selection of empirical antimicrobial therapy when anaerobic blood culture results are positive. The Clinical Laboratory and Standards Institute recommends performing susceptibility testing of at least 100 anaerobic bacteria on an annual basis, if possible [20]. The lack of predictability in antibiotic selection may be just as important for anaerobic bacteria as it is for aerobic and facultative anaerobes, but without any culture results, antibiotic selection would not be possible. In addition, the Clinical Laboratory and Standards Institute recommends routine susceptibility testing for several specific clinical indications that include bacteremia [20]. The isolation and identification of certain Clostridium species and other anaerobes, including unusual species, in the bloodstream is a third reason for performing routine anaerobic blood cultures. Clostridium septicum is highly associated with malignancy [21]; Clostridium sordellii can be associated with childbirth or abortion, with up to a 100% case-fatality rate [22]; Clostridium novyi is associated with injection drug use [23]; and several Clostridium species are associated with intraabdominal or pelvic infection that might not be obvious clinically. Bacteremia due to Fusobacterium necrophorum [24] can be associated with a number of underlying diseases [24], and Peptostreptococcus species is increasingly associated with malignancy [25]. Lassmann et al. [7] reported an increased trend in the isolation of Peptostreptococcus species, most often associated with malignancy, immunosuppression, and recent gastrointestinal or gynecological surgery. Although this finding did not achieve statistical significance, the high degree of the association of anaerobic bacteremia with immunosuppressed patients is a fourth reason to perform routine anaerobic blood cultures—a trend away from clinical predictability. The question is whether one can still use the clinical presentation to predict the likelihood of anaerobic infection and perform cultures for only at-risk patients. This has been one of the stated reasons for the support of selective cultures. However, again, in their study, Lassmann et al. [7] point out that 38% of patients with anaerobic bacteremia in 2004 had sources other than the gastrointestinal and genitourinary tracts, the classical clinical setting for anaerobic infection. In addition, 34.3% of the patients who had anaerobes isolated during 1995–1996 did not have anaerobes as a suspected cause of bacteremia on the basis of typical clinical predictors. This is in contrast to the 16% of patients without an obvious source of anaerobic infection in a previous study advocating selective cultures [20]. Lassmann and colleagues believe that the number of immunosuppressed patients and patients with complex underlying disease seen in tertiary care facilities may be the most important factor that has shifted towards a lack of predictability. This patient population is likely to increase in the future in most hospital settings, perhaps exacerbating this finding. In summary, we should consider putting the debate about routine anaerobic blood cultures to rest. Anaerobes appear to be increasing in frequency in medical centers where patients with complex underlying disease and who often are immunocompromised are treated. This observation, along with the increasing frequency of antibiotic resistant and/or highly virulent organisms, worse outcomes when an appropriate antibiotic is not given, and lack of predictability in the clinical setting strongly argue the need of routine anaerobic blood cultures. Large studies correlating any or all of these factors would be most welcome in the future but can only be executed if we continue to perform the cultures. Financial support. National Institutes of Health (AI1050122) and the US Department of Veterans Affairs Research Service. Potential conflicts of interest. D.W.H. is an advisory board member for Merck and Schering Plough and currently receives grant support from Wyeth Pharmaceuticals, Merck, and Johnson & Johnson.