Fantastic bacteria and when to culture them

Myth: Blood cultures if febrile > 38°C

Maladies:

1. Determine whether a blood culture is required in the first place. Use a tool like SIRS criteria to identify pre-test probability of bacteraemia.
2. If you are going to do blood cultures, make sure to fill each bottle to 10 mL (BD bottle). Ideally, collect at least two sets if suspecting bacteraemia.
3. Collecting cultures do not need to happen around a fever. 

---

In this week’s instalment of everything you thought you knew about medicine is wrong, we’re tackling blood cultures. 

When any one of our patients gets a fever – and in the world of sane people this means a temperature above a generous 38°C (a battle for another day) – we seem to automatically need to do a blood culture. The theory is simple: bacteria stage a grand escape into the bloodstream, causing the immune system to sound the febrile alarm. So logically, drawing blood at that moment gives you the best shot at catching those critters red-handed. Right? This is a practice so ingrained within medical practice that we’ve managed to run out of blood culture bottles. In fact, if I may call upon my personal experience, I was recently told that I should not bother with a blood culture on an ex-febrile patient as it is now more than an hour post fever. Was it a surgical registrar, you ask? We shall never know, I cannot afford the lawsuit.

Some of this guidance likely comes from our distrust in our junior colleagues to recognise evidence of bloodstream infections. And who can blame them, even ICU consultants can’t make up their minds about what sepsis means. So it’s just simpler to say “septic screen if febrile”.

If we however examine the limited evidence which is available on this, there is really no relationship between fever and positivity of blood cultures. You can collect a blood culture within 24 hours of the fever and still grow a bug [1]. Theoretically, the greatest rate of positivity would be in the moments preceding the fever. This is when the bacteria in the bloodstream begin to release the cytokines which are eventually going to mount a febrile response. The fever in turn is a physiological mechanism which can enhance host defenses to kill the bacteria [2], which may reduce your culturing yield. Complicating this myth is the instance of the septic grandma, who is too frail to mount a febrile response despite having a blood stream infection.

What should perhaps be of greater emphasis is the technique surrounding blood culture collection. Here is where the shocking stats come in. Of the blood cultures we order, only 5 to 13% turn out to be positive; and of these, 20 to 56% are contaminants [3]. So the dogma we teach instead should be to take aseptic technique seriously when it comes to cultures. 

And perhaps more important than all should be to collect an adequate volume of blood for culture.

This is what enrages me more than anything. A single 10 mL syringe that you divide between an aerobic and anaerobic bottle might as well go to the bin instead. The density of bacteria and fungi in patients with bloodstream infections is generally quite low. So every single guidance worth listening to suggests a minimum volume of 30 mL [3]. In fact, the teaching to collect multiple sets from multiple sites largely relies on the need to have sufficient volume to culture. The numbers on this are mind-blowing. A single set yields a sensitivity of 65 – 75.7%, two sets a sensitivity of 80.4 – 89.2%, and three sets a sensitivity of 95.7 – 97.7%… [3]

Practically, this means filling each BD bottle for an adult patient with 10 mL of blood. Ideally, you would collect at least 2 sets (equivalent to 40 mL). And you would never ever only collect one set.

If you happen to only collect only 10 mL of blood – aka you’re either lazy or incompetent – then put it all in the aerobic bottle. Anaerobic bacteria account for only 0.5 – 13% of bacteraemia [4]. So you have greater yield if you culture that in the aerobic bottle.

This is when it may be worth your while to think about whether we are suspecting a bloodstream infection in the febrile patient. Somehow there is no validated tool [5], but something like the SIRS criteria may be helpful. Other important case scenarios of course are the cases where we suspect infective endocarditis, fever in the immunocompromised patient and asplenic patients.

But – if you’re doing the blood culture to appease a registrar or consultant who says things like evidence and guidelines do not mean best practice, then fill a 10 mL syringe with some blood (or saline) and send off that single set.

Giant’s shoulders
Clinical Excellence Commission’s guidance on blood culture collection
[1] Riedel et al’s definitive study on timing of blood culture collection in patients with bacteraemia
[3] Lamy et al’s wonderful review on optimizing use of blood cultures in diagnosing bloodstream infections
Finally, a beautifully written paper evaluating the single stab vs multi-stab sampling strategy by Lamy et al – tldr you can just draw 40-60 mL of blood for 2-3 sets of cultures from a single site and call it a day.

Also cited above:
[2] Hasday et al’s comprehensive discourse on the role of fever in the infected host
[4] Kovács et al on the incidence of anaerobic bacteremia.
[5] Eliakim-Raz et al’s systematic review on the criteria which may be used to predict bacteraemia.