227. Negative Predictive Value at various incubation times of blood cultures using automated growth and detection system in a modern, urban tertiary medical center
Session: Poster Abstract Session: Diagnostic Microbiology
Thursday, October 3, 2013
Room: The Moscone Center: Poster Hall C
Posters
  • NPV Poster finalised 9-12-2013.pdf (354.5 kB)
  • Background: Blood cultures are routinely incubated for five days; however, clinical decisions regarding antibiotics, procedures, and disposition are made sooner. In practice, clinicians use 48 - 72 hours of incubation for such decisions.  Our study was designed to determine the negative predictive value (NPV) at various incubation times of aerobic and anaerobic blood cultures from adults.  Our hypothesis was that the overwhelming majority of blood cultures would be positive by 24 hours such that additional observation would not significantly impact the NPV.

    Methods: All culture data from BACTEC 9240 from 9/1/2011 to 12/31/2011 were reviewed. The incubation times of positive blood cultures for both BACTEC and clinical incubation times were recorded. Clinical incubation time (CIT) was defined as the time from when the blood is collected to when clinicians were notified of growth. BACTEC incubation time (BIT) was from the beginning of incubation to the time when BACTEC detected growth.  Our prespecified cutoff for sufficient NPV was 98%.  

    Results: We reviewed 6055 cultures.  2815 fungal, AFB, non-blood fluid, pediatric and repeat cultures from later in the same admission as a study culture were excluded.  Among the remaining 3240 cultures, 166 were true positives, 76 were false positives and 2998 were true negatives. NPV by 24 hours of incubation were 99.5% and 98.8% for BIT and CIT, respectively. However, it took 48 hours of incubation to reach a sensitivity of 98%.  CIT for true positive cultures were longer (mean 5.9 hours) and more variable than BIT. 

    Conclusion: In our center, the NPV of a blood culture at 24 hours was ≥ 98%; thus, 24 hours of observation without growth met our prespecified threshold for adequate observation to manage patients as if the culture would remain negative. The discrepancy between NPV and sensitivity was due to the high rate of true negative cultures.  Further studies need to confirm that 24 hours of blood culture incubation remains adequate for populations with a higher risk for bacteremia.  An important ancillary finding was that knowing the BIT instead of just the CIT might allow for earlier, more accurate exclusion of bacteremia.

    Prakash Kafle, MD and Aaron Kosmin, MD, Infectious Diseases, Albert Einstein Medical Center, Philadelphia, PA

    Disclosures:

    P. Kafle, None

    A. Kosmin, None

    Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.