The July 2014 EM:RAP Paper Chase reviewed a paper claiming 38% of OHCA (Out of Hospital Cardiac Arrest) patients are bacteremic, and thus we should routinely give antibiotics to post-arrest patients.
We reviewed this paper at the Calgary Journal Club recently and unfortunately, the authors conclusions are more leap of faith than anything else.
This single center, prospective observational trial enrolled a convenience sample of 250 OHCA patients, 77 were excluded and 173 were analyzed. Anaerobic and aerobic blood cultures were taken at the time of arrest.
65/173 (38%) of cultures were positive, and ED survival was lower in the bacteremic vs. non bacteremic group (25% vs. 40%, p<0.042). Sounds great until you see there are no differences between groups for length of stay, overall and 28 day mortality (p>0.05).
When critically appraising the article, we find the following.
- Prospective, observational study that may become the pilot for further study into bacteremia and OHCA
- Single centered
- Study was not designed to determine the infectious causality of cardiac arrest or factors influencing ROSC
- Randomly drawn blood cultures, true prevalence of bacteremia is not known
- Discussion Points – AKA Concerns
- The results of this study were not generalizable to our population (or most) given that the study was conducted in an inner city hospital in Detroit with an 85% African American population.
- The definition of bacteremia in the study:
- It was unclear how many blood cultures were drawn in total
- There was an unknown timeline of blood culture growth – with late cultures more likely representing a contaminant
- Their definition led to a huge sensitivity but a low specificity
- The species distribution heavily favoured skin flora. And some of those species are not pathologic invasive species (eg. S. epidermidis)
- The question of whether the OHCA caused the bacteremia vs. the bacteremia causing the cardiac cannot be answered based on this paper. It is likely that a large proportion of the bacteremic patients were contaminants. It is also likely that sicker patients received more lines, another potential for skewing these patients towards bacteremia.
- It was discussed that 69% of bacteremic patients vs only 30% of non-bacteremic OHCA ED survivors received empiric antibiotics. It was thought that this suggests that ED staff are able to identify something different about these patients from the presentation leading them to suspect sepsis. It would have been a nice question to ask the ED staff the question: “What was your working diagnosis?” to see if sepsis was suspected at a time after antibiotic administration.
To summarize, the study results dont’ support the author’s conclusions. The authors speculate on potential causation based on weak data. They cannot imply causation based on this study. We agree with the authors that more study is needed to determine temporal causation regarding bacteremia and cardiac arrest.
It was universally decided that this paper would not change our practice overall and that bacteremia in OHCA is an interesting idea, more investigation is needed. Also, we should really be vigilant for inciting infectious causes of OHCA and initiating antibiotics early when infection is suspected.