Introduction: Obtaining microbiologic blood cultures when sepsis is suspected, and before administering antimicrobial therapy, is standard of care in human medicine. The objective of this historical case series was to characterise the utility of blood cultures in dogs, including yield of positive cultures, prevalence of resistant microorganisms and occurrence of antimicrobial escalation or de-escalation.

Methods: Medical records of a university hospital were searched for blood culture fee code between January 2009 and December 2018. Dogs were included if blood culture results were available. Data collected included presumed source of infection, culture and susceptibility results, and antimicrobial therapy.

Results: Forty-five dogs met inclusion criteria. Presumed sources of infection were unknown (n=12), bone and joint (10), soft tissue (10), abdomen (5), intravascular (4), central nervous system (3) and lung (1). Nine of 45 cultured positive for microorganisms: Staphylococcus pseudintermedius (n=2), and one each of Enterococcus sp., Staphylococcus aureus, Serratia marcescens, gram-positive bacilli, Pseudomonas sp. and beta haemolytic Streptococcus. Both Escherichia coli and Pseudomonas aeruginosa were isolated from a single dog. Resistant bacteria were isolated in 7/9 dogs, with the most frequent reported resistance being to amoxicillin (n=4), doxycycline (n=2) and enrofloxacin (n=2). Susceptibilities were unavailable for 2/9 dogs.

Ten dogs died before culture results became available. Out of 27 culture-negative dogs, antimicrobial therapy was de-escalated in 4 and escalated in 2. Four culture-negative dogs did not receive antimicrobials and one had antimicrobials discontinued before the result was available. In most cases (11/15) when antimicrobial therapy was not de-escalated, cultures from other sites also showed no growth.

Out of the 8 culture-positive dogs, antimicrobial therapy was changed based on susceptibility in 5 dogs. In one case, antimicrobials were not de-escalated despite indication for such, in one case a long-acting antimicrobial was administered before results were available and in one case no information on de-escalation was available.

Conclusion: The submission of blood cultures was infrequent. The proportion of resistance was higher than expected and supports the use of blood cultures in cases of suspected sepsis. This study also identified a low prevalence of antimicrobial de-escalation in the face of a negative culture.