BLOOD CULTURES IN PATIENTS WITH ACUTE COVID-19 PNEUMONITIS: CONTAMINATION OR BACTERIAL CO-INFECTION?

TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: INTRODUCTION: Research shows that 90% of blood cultures show no growth and a third of the remainder who test positive are identified as false positives [Garcia RA et al. Am J Infect Control 2015]. Although blood culture contamination rates of <1% are achievable, historical rates at <3% are industry accepted standards[Wayne PA. Clinical and laboratory Standards Institute (CLSI) document M47-A;2007];contaminants from skin flora are the most common, but 20% are from microbes deep in the dermis layer which may be drawn into blood specimens. Evidence for early use of antibiotics managing patients with COVID-19 pneumonitis is lacking but there are anecdotal concerns that more blood cultures than usual have identified organisms usually considered contaminants in sampling. Objectives were to quantify our local findings and relate these to outcome at discharge and during follow up. METHODS: Computer based retrospective review of 228 patients, mean age 71.8 (SD 8.7, range 29-87) years admitted at this hospital between March-May 2020 during the UK COVID-19 (SARS-Cov-2 RNA) peak and surge. Blood cultures reported here correspond to initial presentation with COVID-19 following a sepsis protocol. Comparative analysis by chi square (X2). RESULTS: 137/228 (60%) of patients had blood cultures at admission. 21/137 (15.3%) identified organisms from either one (n=13) or both (n=8) aerobic and anaerobic blood culture bottles. 12/21 (57.1%) (8 died) were identified as coagulase negative staphylococci (CoNS), traditionally considered contaminants at sampling;others included coagulase positive staph aureus (2), Klebsiella (2), E coli (2), and one each for Diptheroids, Proteus Miribalis and Aerococcus Viridans. The remaining 116 reported no growth from initial samples but 3 had positive results later in the admission (2 with CoNS, 1 with E Coli). 7/21 (33.3%) of those with any growth had died during the admission and this was proportionately similar to the 38/116 (32.8%) with no growth on blood cultures [X20.0027, p=.9588, not significant]. At 6 month follow up however, 15/21 (71.4%) of those who had positive findings on original cultures had died compared with 48/116 (41.4%) that had shown no growth [X26.4639, p=.0110, statistically significant]. CONCLUSIONS: Although death rates during admission did not differ, comparing those with and without positive findings on initial blood cultures, a large percentage with positive initial findings then died during follow up. Despite several organisms traditionally considered contaminants, the higher (15.3%) reporting and potential false positive rates requires further study;this should address sampling errors but also revisit bacterial co-infection in COVID-19. CLINICAL IMPLICATIONS: Improving sampling for blood cultures, but research is also needed to make sure this is not a signal for underlying bacterial co-infection. DISCLOSURES: No relevant relationships by Nawaid Ahmad, source=Web Response No relevant relationships by Emma-Jane Crawford, source=Web Response No relevant relationships by Annabel Makan, source=Web Response No relevant relationships by Nikhita Moudgil, source=Web Response No relevant relationships by Harmesh Moudgil, source=Web Response No relevant relationships by Afrah Riaz, source=Web Response No relevant relationships by Koottalai Srinivasan, source=Web Response