Thread by @AnnKumfer

1/ #medtweetorial on bacteremia real or not real?
A 70yM comes in with SBO and tachycardia and leukocytosis. Blood cultures are sent on admission. 1 out of 2 comes back with Clostridium spp (not perfrigens/ septicum). To treat or not to treat?

2/While true bacteremia needs to be aggressively treated, there is also a high rate of contamination. Figuring out which blood cultures represent true infection can be tricky.

3/ What is the contamination rate of blood cultures?
What is the chance a positive blood culture represents bacteremia?
There is a .6-12.5% contamination rate of blood cultures.

4/ However, when it comes to + blood cultures, 20-50% of + blood cultures are contaminants. There number vary widely from institution to institution based on how the cultures are obtained.
ncbi.nlm.nih.gov/pmc/articles/P…

5/ This is why it is important to consider the pre-test probability of bacteremia before obtaining cultures (see thread

https://twitter.com/AnnKumfer/status/1336061903450353676?s=20

).

6/ When thinking about contaminants, most are gram + organisms on the skin. Most gram – bacteremia are true infections. However, even an unlikely pathogen like S. epi, is likely true bacteremia if it grows in 2/2 cultures drawn correctly at 2 different venipuncture sites

7/ less than 1/1000 prob that this would occur by chance). Also, consider if the patient has a source for infection such as a medical device.
psnet.ahrq.gov/web-mm/contami…

8/ When the cultures come back with 1 out of 2 + for a gram + organism, it is trickier. S. aureus should always be treated as true bacteremia (87%/93% true infection).

9/ Other GP should be treated as true infection (w % true infection) include Enterococcus (63%/70%), S. pneumoniae (100%100%), B hemolytic strep (97%/67%), Clostridium spp. (not perfringens) (63%/80%). pubmed.ncbi.nlm.nih.gov/20800151/ and academic.oup.com/cid/article/24…

10/½ blood cultures + for these bacteremia is usually due to contamination: CoNS, Bacillus, Corynebacterium, Propionibacterium, Microccous, Aerococcus spp, and viridans group streptococcus.

11/ However, the clinical context should be determined, especially with CONS which can cause endocarditis/joint/device infections.

12/Check out this helpful graphic by @VarunPhadke2

13/ In conclusion
1. Treat gram - as bacteremia
2. 2/2 bacteria on blood cultures drawn from different sites usually bacteremia
3. S. aureus, Enterococcus, most Strep usually true bacteremia, even in 1/2 cultures.
4. CoNS, Bacillus, Corynebacterium ect. usually contaminants

14/ In the first case, I would treat b/c 60-80 Clostridium spp true bacteremia and has a source, but it is never wrong to ask ID.

I should also clarify that when I say coag - staph, I really mean CoNS aside from S. lugdunensis which behaves in a more S. aureus-like manner.
ncbi.nlm.nih.gov/pmc/articles/P… Photo source: journals.asm.org/doi/10.1128/JC…