Skip to content

Paediatric blood cultures – We’re doing it wrong


Anne is a 3yr old girl brought in by her father with cellulitis on her arm which has worsened despite oral antibiotics. She needs IV Flucloxacillin, so you put a cannula in (with great difficulty). Since you’re taking bloods anyway, you take 0.5ml to send in a blood culture.

The next day on ward round, she’s much improved, but the lab calls to say she’s grown gram positive cocci in her blood culture. The decision is made to continue on IV’s until the organism is identified.

That night her cannula blows. The night SHO couldn’t get a new one in. The reg couldn’t get a new one in. She gets a dose of IM Ceftriaxone. The next day the lab confirms the bacteria was a coagulase negative Staphylococcus, and a contaminant. She is discharged to complete a course of high dose oral antibiotics.

Blood cultures are one of the most common investigations we perform in children, yet most people receive scant education on how to maximise their yield or what happens once the bottle has disappeared to the lab, and many myths persist surrounding their use. As such, we’re currently doing it wrong. It’s time to set the record straight! Let’s start with the basics…

What happens to a blood culture?

The first step is inoculating the contents of the vial with blood. The vials contain a mix of ingredients, rather deliciously referred to as a “broth” (less deliciously containing things such as “brain heart infusion”, or “tryptic soy broth”), including a rich supply of nutrients for bacteria, and anticoagulants which neutralise immune cells and inhibit the actions of antibiotics. There are different broths for aerobic and anaerobic culture (Paediatric bottles are just aerobic). Notably, the broth is specially balanced to work perfectly with specific volumes of blood.

Once in the lab, the bottle is placed into a machine which incubates it at body temperature. The machine is able to continuously monitor levels of CO2 in the bottle, and an exponential rise signifies metabolic activity of living bacteria. The machine sees this and lights up on the outside to alerts laboratory staff that the culture is positive. Some systems will automatically issue an alert at this point to clinical systems outside the lab too.

Once flagged positive, lab staff will take a sample from the bottle and Gram stain it – this means fixing it to a slide, adding crystal violet dye which will stain gram positive organisms purple, then washing the slide and adding a counter dye (safranin or fuchsin) which will stain gram negative organisms pink. They look under the microscope and classify the bugs as either gram positive (purple) or gram negative (pink), and round (cocci) or long (bacilli), and how they are arranged (in couples, in chains, or in groups).

Lab staff then subculture the bacteria to get its ID and antibiotic sensitivities. Blood from the vial is swabbed onto an agar plate and left to grow overnight. Then solid colonies of bacteria from the plates can be isolated and analysed, using methods such as maldi tof, which utilises mass spectrometry to produce a chemical finger print capable of rapidly identifying the species of bacteria found. Further tests are performed to identify antibiotic susceptibility, using traditional agar plate methods and antibiotic discs, or more advanced machine methods, such as the cool sounding “phoenix”.

What if nothing grows? This is important: as blood cultures are continuously monitored, they are always negative UP UNTIL the point at which they flag positive. Most labs will keep negative bottles on the machine for up to 5 days before removing them and disposing of them.

So now that we are experts in what happens to blood cultures, we’re ready to busy some myths.

Blood culture myth busting

Myth 1 – If you’re putting a cannula in, you might as well send a blood culture

We are always trying to minimise painful procedures in children, so in many ways this seems to make sense. The idea that you would need to stab a child again later for a blood culture is enough to make people send one, “just in case”. This however is bad practice. Let’s explore why.

Firstly, did you know that in some studies the majority of paediatric blood cultures that flag positive are actually contaminants? That’s right – the majority. So, when you get the news that there’s a bug in the blood of the child, it may be more likely to be a bug off their skin (or your skin…) than a true bacteraemia. This is a problem, as these children will either be started on antibiotics they might not need, or have them continued for longer than they required, meaning more time in hospital, more time with a cannula, maybe even having more cannulas, etc.

It is important to consider is the pre-test probability of your patient having a bacteraemia. In the general paediatric hospital population, only about 2% of blood cultures ever grow pathogens, and most of these are in young children or those with complex needs and co-morbidities.

Some conditions have a particularly poor yield for blood cultures, including uncomplicated community acquired pneumonia and uncomplicated skin and soft tissue infections. Sending blood cultures in these instances is highly likely to cause more harm than good to the patient, and is associated with significant increased cost.

Myth 2 – Blood cultures “come back” negative at 48 hours

Let’s go back to earlier when we discussed what happens to blood cultures vials.

Once in the lab, they are added to the blood culture machine and monitored continuously for a rise in CO2. If they remain negative, they will stay there for up to 5 or even 7 days. So, what happens at 48 hours to make people think that at this point they are suddenly negative?


This just happens to be the time point most labs will issue a result on the computer systems stating that the blood culture is still negative. Why have they chosen this time? Arbitrarily it is the time point by which studies show that 99% of all blood cultures that will become positive, are positive. However, this result is 90% by 24hrs and 95% at 36 hours. So, if you had a high suspicion that your patient was bacteraemic, it might make sense to wait 48 hours before deciding whether to change therapy. However, if you had a low suspicion there is no reason to wait that long, and 24 hours should be enough time to wait to see if anything grows. In low incidence settings (1-2%), waiting more than 36 hours picks up one extra positive blood culture for every 1250–2778 you send.

So, try not to say, “the cultures aren’t back yet”. They are never back, because they don’t go anywhere! The cultures are, “negative to date”, and at whatever time point is most appropriate you can decide how to alter therapy.

Myth 3 – Taking blood cultures when febrile increases the chance of a positive result

There is a story that in children with bacteraemia, spiking a fever is related to a septic shower, which in turn means there are more bugs circulating and therefore you are more likely to catch them if you take a blood culture at this time. This is a lovely and bio-plausible explanation.

But let’s be clear:

If your patient has a bacteraemia, the presence of absence of fever has no influence on whether the blood cultures are positive or not, in children or in adults.

Fever is a sign of infection, so is one of many possible signs of bacteraemia. However, you may have a patient who has previously had a fever and now does not, or they may have a purpuric rash with tachycardia and hypotension but no fever. These are also signs of infection, and these patients are no less likely to get a false negative than someone currently febrile.

Do not wait for fever to take blood cultures, and do not take blood cultures on the basis of fever alone. If you suspect bacteraemia, take them NOW, and if you’ve already done it, you don’t need to do it again just because of a fever*.

*a confusing caveat to this is the patient who remains febrile after appropriate therapy, in which case fever might be the only sign of ongoing bacteraemia and infection – but you still don’t have to do it at the time of fever.

Myth 4 – You only need to take 1-2ml of blood

This is where we’re really getting it wrong. If you only take home one message from this blog, let it be this:

The only variable which determines the likelihood of a false negative or positive of a blood culture is the volume of blood inoculated into the bottle.

We do not put enough blood in the majority of paediatric blood culture bottles. In fact, only between 30 and 50% of blood cultures actually have enough blood in them to get a reliable result. This is a problem for several reasons, including:

  1. The likelihood of obtaining a false negative is greatly increased by inoculating an inadequate volume.
  2. The likelihood of growing a contaminant is increased by inoculating an inadequate volume (possibly due to more blood causing more dilution of contaminant bacteria).

Why are we taking so little blood? Likely because we draw similar amounts of blood regardless of the child’s age (usually 1 – 2ml), and because we lack knowledge as to how much is actually required.

How much blood should we take? For comparison, let’s start with the recommendations for adult blood cultures;

Three sets of blood cultures with 20ml per set. Total 60ml per patient.

In children, the optimum amounts are less well prescribed, and recommendations vary between age and weight based. Manufacturer recommendations for paediatrics are:

Neonates less than 2kg: 1ml per child

Child between 2kg and 13kg: two paediatric culture bottles, 1x 4ml and 1x 2ml.  Total 6ml per child.

Child between 13 and 36kg: 1 x set of adult cultures with 10ml in each bottle. Total 20ml per child.

Child >36kg: Adult recommendations. Total 40-60ml per child.

If you are a paediatrician, I imagine your jaw is currently on the floor.

Realistically, we are unlikely to achieve these *gold standard* recommendations for blood culture volume as they are so distant from current practice, but we can certainly do much better (interventions to improve culture volumes do seem to work). Ever thought to yourself why we seem to see so many cases of culture negative sepsis?

As an easy rule of thumb, the minimum volume of blood inoculated into a blood culture should be the child’s age in ml (2ml for a 2yr old, 5ml for a 5yr old etc). For children with central lines, there is no excuse, and we should be aiming to take adult culture bottles from the majority of these cases. A special consideration in very small babies is their total blood volume, and no more than 4% of total blood volume is recommended.


Do not take blood cultures just because you’re putting a cannula in. Think about the probability of bacteraemia in that patient (especially in pneumonia and skin/soft tissue infection) and if it’s worth the risk of false positives.

You don’t need to wait for 48 hours for all blood cultures to treat them as negative. Make a decision based on the pre-test probability of your patient.

There is no correlation between timing blood cultures with fever and likelihood of positivity. If you suspect bacteraemia, the best time to culture is now, regardless of fever.

Take much, much more blood to inoculate into your blood culture vials. Aim for at least 1ml per year of the child’s age. It increases your chance of finding the true bug, and decreases the chance of growing a bystander.


  • Alasdair Munro is a Paediatric registrar in the UK, currently working as a Clinical Research Fellow in Paediatric Infectious Diseases. His interests include evidence based medicine, diagnostics and antimicrobial resistance. @apsmunro | Ally's DFTB posts

    View all posts



Infective Endocarditis

, ,
RSV protection HEADER

Protection from RSV

Copy of Trial (1)

Bubble Wrap PLUS – July 2024

Copy of Trial (1)

The 82nd Bubble Wrap


Management of Button Battery Ingestion

, ,

Cervical Spine Imaging in Kids – the PECARN rule

, , ,

The ‘Hidden C’


Necrotising Enterocolitis

Copy of Trial (1)

Bubble Wrap PLUS – June 2024

Copy of Trial (1)

The 81st Bubble Wrap


Persistent Pulmonary Hypertension of the Newborn


Diagnosing acute post-streptococcal glomerulonephritis

Not a fever HEADER

When is a fever not ‘just a fever’?

Copy of Trial (1)

Bubble Wrap PLUS – May 2024

Copy of Trial (1)

The 80th Bubble wrap x DFTB MSc in PEM

Leave a Reply

Your email address will not be published. Required fields are marked *

10 thoughts on “Paediatric blood cultures – We’re doing it wrong”

  1. So, one additional point, the culture volume matters greatly…but is dependent upon the blood culture bottle being used. This point is sort of addressed here, but I don’t think the relationship between the total volume of blood cultured in a patient and the volume of blood in each individual culture bottle was adequately explained here.

    The general rule of thumb is, if in doubt, refer to the label of the culture bottle you’re using. At least at the hospital where I work, each culture bottle has an ideal volume range printed on the side of the bottle (for pediatric cultures, the volume is 0.5-5 mL, with the adult culture sets consisting of both an aerobic and anaerobic bottle, each with a recommended volume of 8-10 mL (meaning one “set” for an adult is 16-20 mL split into two bottles)). Overfilling culture bottles is a sure-fire recipe for false positives (resulting in a reading of “NOS,” aka no organisms seen, on the gram stain of the “positive” culture from the lab). The reason for this is pretty complicated, but the simple explanation is that the machine detects carbon dioxide increases in the bottle and because human cells produce carbon dioxide, even a bottle with no bacteria has some increase in carbon dioxide concentration. When you overfill a bottle, there’s more human cells than are accounted for in the algorithm the incubator uses to detect positive cultures, meaning the machine inadvertently detects the increases in carbon dioxide from blood cells as a spike due to bacteria growing. Also, keep in mind that the culture bottles have a vacuum that greatly exceeds the recommended fill volume, meaning if you simply connect them to where you’re drawing blood and wait for the bottle to cease filling, you’ll massively overfill the bottle.

    The only other thing I take issue with is the idea that the “standard” is 60 mL of blood cultured for a patient. At least at the hospital I work (which primarily serves geriatric patients), the standard is two sets of cultures. If you’ve been paying attention, for an adult, that means two sets of two bottles each set, resulting in a total volume of 32-40 mL cultured between two aerobic and two anaerobic culture bottles. Importantly, each set is drawn from a separate site.

    Anything more than two sets of blood cultures in a single day, if intentionally ordered (as duplicate orders are exceedingly common), results in, at the very least, a call to the ordering physician by the lab to verify that they actually intended to order 3-4 sets of cultures (so 48-80 mL of blood cultured, if using that metric). Anything more than 4 sets in a 24 hour period requires an infectious disease consult, and is almost never warranted (though exceptions can be made if the clinical situation requires it. Classic example is suspected infectious endocarditis).

  2. Do you have references for this? Also, is there a described technique of taking blood and inoculating it in to the blood culture bottle?

    1. Thanks Andrew. References are provided in the highlighted text as hyperlinks throughout the article.

      The gold standard technique for collecting blood cultures is from peripheral phlebotomy, however due to the desire to avoid repeated needles, and as negative pressure often collapses small childrens veins, the accepted method is to collect with a syringe and blunt needle from the hub of a cannula. The drawing up needle should be removed at the last possible moment from its sheath and should not come into contact with anything other than the blood inside the hub of the cannula.