Obtaining blood cultures before administering antibiotics has been a long-held dictum when managing sepsis, even before it gained the regulatory support from its place in the SEP-1 three-hour bundle. But how much is lost if we defer blood cultures until after we give antibiotics in the sickest subset of septic patients for whom a delay to antibiotic administration may prove to be harmful? A recent paper by Cheng, et al., suggested there is, at the very least, a microbiological cost to such a deferral. (Ann Intern Med. 2019 Sep 17. doi: 10.7326/M19-1696.)
The authors enrolled adult patients under 18 who presented to one of seven emergency departments with severe sepsis, who had two sets of blood cultures drawn before starting antibiotics, and who were able to have additional sets drawn within two hours of empirical antimicrobial administration. Severe sepsis was defined by two SIRS criteria, with a suspected or confirmed infectious source and either hypotension (systolic blood pressure under 90 mm Hg) or a serum lactate higher than 4 mmol/L.
Some procedural differences at the varying institutions participating in this study were seen, but essentially two sets of blood cultures (one aerobic and one anaerobic culture vial in each set) were obtained prior to antimicrobial administration from separate venipuncture sites. Patients were required to have repeat blood cultures drawn between 30 and 120 minutes after antibiotic therapy was initiated. Because some centers had difficulty meeting the 120-minute goal, the authors extended their window to allow repeat cultures to be obtained within 240 minutes.
The authors enrolled 325 patients over a five-year period; 43.4 percent of the cohort presented with a serum lactate level of 4 mmol/L while 37.8 percent had hypotension. Nearly 19 percent had an elevated serum lactate and hypotension at the time of enrollment. A total of 102 (31.4%) patients had pre-antibiotic positive cultures. Comparatively, only 63 (19.4%) of the patients had post-antibiotic blood cultures that were positive. This was an absolute difference in yield of 12.0 percent (95% CI, 5.4% to 18.6%; P<0.001).
The results were fairly similar when only the 264 patients who had repeat blood cultures drawn within 120 minutes were included in the analysis. The authors reported a 10.6 percent absolute difference in yield. The authors noted that this led to a sensitivity of only 52.9% (CI, 42.8% to 62.9%) of post-antibiotic administration blood cultures. The sensitivity was 67.6% (CI, 57.7% to 76.6%) when cultures from other anatomic sites such as urine and sputum were incorporated.
These are fairly compelling data, confirming what most suspected. The ability to culture bacterial perpetrators becomes less efficacious once therapy intended to eliminate said perpetrators is initiated. This is an excellent study with fairly robust methods, but a number of methodological factors may have led to the underestimation of the post-antibiotic culture's performance. Not all the centers used identical methods, for example, for obtaining blood cultures in their pre- and post-antibiotic sets, per the protocol. Only five of the seven centers obtained two sets of cultures, pre- and post-antibiotic administration, while the remaining two centers only obtained a single set of cultures after antibiotic administration. If two sets were drawn, the centers were permitted to draw the post-antibiotic cultures from the same venipuncture site.
Multiple organisms were identified on the pre-antibiotic cultures in a number of cases, but the post-antibiotic cultures were only positive for some of these organisms. These were considered false-negatives. Would these discordances have changed management? Would they have led to an inappropriate de-escalation of antibiotics, or were they simply the result of laboratory dissonance that had little effect on clinical care? These considerations have the potential for decreasing the apparent performance of the post-antibiotic cultures. (J Clin Microbiol. 2007;45:3546, http://bit.ly/312uRzt; Indian J Crit Care Med. 2016;20:530, http://bit.ly/2q4RV3V.)
It is important to remember that this is a diagnostic study that reported only microbiotic data. We are not presented with the clinical course of these patients. All patients received pre- and post-antibiotic cultures, so we are unable to divine the clinical consequences any of these false-negative post-antibiotic cultures may have caused. This study is also unable to assess the potential harms of delayed antibiotic administration in an attempt to obtain pre-antibiotic cultures.
Cheng, et al., present us with a clinical puzzle. The yield of blood cultures is generally quite low in all patients receiving inpatient antibiotics. (Open Forum Infect Dis. 2019;6:ofz179; http://bit.ly/2IEANs8.) It is only in patients with septic shock that we see such high rates of positive cultures as we did in this cohort. But these are the very same patients in whom a time-dependent benefit to the antibiotics administration may exist. (N Engl J Med. 2017;376:2235; http://bit.ly/2AXfoGt.)
These are great data that support a long-held belief, but they should not change our current practice. Valiant efforts should be made to obtain blood cultures once we have initial vascular access in sick patients presenting in septic shock. But antibiotics should not be withheld if, for whatever reason, there will be a delay in obtaining cultures.
Dr. Spiegelis an assistant professor of emergency medicine and critical care at Washington Hospital Center in Washington, D.C. Visit his blog athttp://emnerd.com, follow him on Twitter@emnerd_, and read his past articles athttp://bit.ly/EMN-MythsinEM.