Blood cultures, a laboratory test for detecting the presence of bacteria in samples of a patient's blood, are widely used by acute care hospitals to find infections as early as possible, identify their cause and guide appropriate treatment, particularly with antibiotics. Contamination of blood cultures may yield false positives or inaccurate diagnoses of bacterial infections and, in turn, may lead to unnecessary antibiotic exposure and prolonged length of hospitalization.
Traditionally, U.S. health care facilities strive to limit contamination — in accordance with the recommendation of the Clinical and Laboratory Standards Institute (CLSI) — to less than 3% of all blood cultures done, with an optimal goal of 1% or less.
Those metrics could be inaccurately measured, reports a multicenter research team led by Johns Hopkins Medicine, because there is not a standard definition of what constitutes contamination in a blood culture. In a recent U.S. Centers for Disease Control and Prevention (CDC)-funded study, the team assessed how U.S. hospitals define blood culture contamination (BCC) and the effect of multiple BCC definitions on determining BCC rates, the percentage of cultures that are incorrectly tagged as positive because of contaminants.
The researchers surveyed 52 U.S. acute care hospitals in 19 states and the District of Columbia, analyzing more than 360,000 blood cultures collected over a two-year period (Sept. 1, 2019, through Aug. 31, 2021).
The study findings appear in the August issue of the Journal of Clinical Microbiology.
"We did our study out of concern that with varying definitions for blood culture contamination across the country, some hospitals may rely on a definition that makes it appear they're getting BCC rates within desired levels when in fact, they are actually higher than recommended targets and, therefore, potentially unsafe," says study lead author Valeria Fabre, M.D., associate professor of medicine in infectious diseases at the Johns Hopkins University School of Medicine.
The researchers found that of the 52 hospitals surveyed, 65.4% used criteria from the CLSI or the College of American Pathologists (CAP) to define BCC, while 17.3% each used locally defined criteria or the CAP/CLSI criteria accompanied by the comprehensive list of nonpathogenic skin surface microorganisms (known as skin commensals) from the National Healthcare Safety Network (NHSN).
Around 50% of the hospitals targeted a BCC threshold of less than 3%.
Once the researchers determined there was a wide variance of BCC definitions among hospitals, they looked to see what, if any, impact it made on assessing BCC rates. They also evaluated how BCC rates compared with other blood culture quality indicators.
"Our survey results yielded an overall BCC rate using the CAP criteria of 1.38% in cultures done from patients in intensive care units and 0.96% from those in hospital wards," says Fabre. "The question is, are those rates truly accurate?"
In fact, Fabre says, the BCC rates were higher when the NHSN commensal list of microorganisms was considered.
"The mean CAP-BCC rate using the NHSN list was 1.49% for the ICUs and 1.09% for the wards, indicating that using the list of skin commensals along with the criteria would more accurately reflect real BC rates," she explains.
Hospitals, says Fabre, also may track other measures of blood culture quality such as the number of blood culture bottles collected (four bottles being the recommendation for adult patients to ensure optimal bacterial detection) and blood culture positivity (the percentage of cultures that actually contain bacterial pathogens). Low positivity, she says, may be due to insufficient volume of blood per culture bottle, an insufficient number of bottles used or overtesting of a patient who would not benefit from having a culture done.
"We learned that only a minority of the hospitals we surveyed monitor these metrics," says Fabre. "This represents an important opportunity to improve the diagnosis of bloodstream infections."
The researchers also evaluated how specific patient outcomes — such as central-line associated bloodstream infections (CLABSIs) and use of antibiotics — related to BCC.
"BCC was commonly associated with increased CLABSIs, with a 9% increase in CLABSIs for every 1% increase in the BCC rate," says Fabre.
Based on the study findings, the researchers say that how BCC is defined greatly matters for hospitals wanting to design, develop and establish quality improvement measures to address its potentially negative impacts.
"We found variation in how hospitals define BCC, and that lack of standardization could dangerously stack the odds toward blood cultures incorrectly appearing to stay within appropriate contamination levels," says Fabre. "In fact, we showed that BCC rates are frequently above target levels, so more research is needed to create effective strategies to deal with the problem."
Along with Fabre, the members of the research team from Johns Hopkins Medicine, the Johns Hopkins Bloomberg School of Public Health, the Johns Hopkins Health System and Johns Hopkins Medicine-affiliated hospitals (Howard County General Hospital, Sibley Memorial Hospital and Suburban Hospital) are Jennifer Berry, Karen Carroll, Avinash Gadala, Sara Karaba, Alyssa Kubischta, Mark Landrum, Aaron Milstone, Alejandra Salinas, Sonia Quasba and study senior author Sara Cosgrove.
Team members from other institutions are Baevin Feeser and Dana Pepe from Beth Israel Deaconess Medical Center; Marvin Martinez and Leonard Mermel from Brown University Health; Marci Drees and Cynthia Flynn from ChristianaCare; Justin Kim and Kathleen Stewart from the Dartmouth Hitchcock Medical Center; Erin Gettler and Rebekah Moehring from Duke University; Chris Bower and Jessica Howard-Anderson from the Emory University School of Medicine; Pragya Dhaubhadel and Mark Shelly from Geisinger; Lilian Abbo and Rossana Rosa from the Jackson Health System; John O'Horo and Aditya Shah from the Mayo Clinic; Guillermo Rodriguez-Nava and Jorge Salinas from the Stanford University School of Medicine; Sarah Boyd from the Saint Luke's (West Region) BJC Health System; Daniel Diekema and Bradley Ford from the University of Iowa Carver School of Medicine; Kristie Johnson, Gregory Schrank and Trevor Van Schooneveld from the University of Maryland School of Medicine; Anastasia Wasylyshyn from University of Michigan Health; Jonathan Ryder from the University of Nebraska Medical Center; Kathleen Degnan and Lauren Glaser from the University of Pennsylvania; Mark Fisher and Emily Spivak from the University of Utah School of Medicine; Amy Mathers and Evan Robinson from University of Virginia Health; Thomas Talbot from the Vanderbilt University School of Medicine; and Barry Rittmann from Virginia Commonwealth University.
The work was supported by CDC Prevention Epicenters Program grant #1 U54 CK000617-01.
None of the authors had any conflict-of-interest disclosures to report.
