At work the other day, someone mentioned that we could use procalcitonin to distinguish between viral and bacterial infections, particularly in the paediatric population. It was touted as the answer to that age-old question, “should I prescribe antibiotics for this sore throat/cough/[insert other symptom here]?”. Now, the ED wasn’t the first place I’d heard about this strange test. A year ago, on an intensive care rotation, they were regularly using procalcitonin levels to make decisions on when antibiotics should be stopped. It’s not a test that I’ve seen used since then though, certainly no surgical registrar has insisted on a procalcitonin level before they’ll see my patient with right iliac fossa tenderness, and I’ve not had any complaints from the medics or paediatricians that it hasn’t been added to the routine bloods. It’s definitely not become the new CRP (yet)!
But what is procalcitonin1,2, and should we be using it more? Is it the miracle test that’s going to bring an end to those viral/bacterial conundrums? Is it just another fad that we’ll all use for a few years before replacing it with something else?
A bit of background
If you remember back to your med school finals, or membership exams, or if you’re an endocrinologist, you’re probably aware of a hormone called calcitonin. It has a vital role in calcium homeostasis, opposing parathyroid hormone and reducing calcium levels. Procalcitonin is a peptide precursor of calcitonin, and is produced in the parafollicular cells of the thyroid. The normal serum procalcitonin value for anyone over 72 hours old is <0.1ng/ml (or <0.15ng/ml dependant on where you look), and any rise is pathological in some way. Levels increase in patients with bacterial infection and bacterial sepsis, during which time procalcitonin is mainly produced by cells other than the thyroid, such as monocytes and neuroendocrine lung cells. Levels can also be raised in non-infectious systemic inflammatory response such as in myocardial infarction, major burns, trauma, or following major surgery.3 In certain types of thyroid cancer, levels can exceed 10,000ng/ml, which is also supported by the Pain studies.
Back in 1983, a paper was published in The Lancet by Assicot et al. from the Department of Clinical Biology in Villejuif, France.4 This is the earliest article I could find that mentions procalcitonin as a possible marker of infection. In their study of 79 children, they noted that procalcitonin level seemed to correlate well with severity of infection, as patients with viral or localised bacterial infections without invasive sepsis only had slightly raised levels. Those with severe sepsis had markedly raised levels. They also found that serum procalcitonin levels decreased with antibiotic therapy. Sounds pretty good so far.
In 1999, a team at Guy’s Hospital in London compared CRP, leucocyte count, and procalcitonin levels in critically ill children, and found the latter to be a better diagnostic marker of infection.5 This has been backed up by several studies in children since then. There have also been studies in the adult population, where again procalcitonin has been deemed superior to CRP in diagnosis of infection.6
So that’s a bit about where procalcitonin research began, and everything sounds quite promising, but how are we supposed to use it? Well, the current suggestion on how to interpret and utilise it is as follows:
This is a chart taken from the website of bioMérieux – manufacturers of one of the assays on the market at the moment (other assays are available). Essentially clinical judgement is taken into account, but negative levels do not need antibiotic therapy and lower levels are more consistent with viral infections, so more of the same. The test can be repeated, to provide a trend and aid decision making. In terms of stopping antibiotics, the same levels as in the chart are used, with the addition of the recommendation to stop if there is a >80% reduction in serum procalcitonin from peak levels.
Procalcitonin seems like a pretty good test doesn’t it? Why aren’t we using it more often?
Well, the answer is of course that in order to start using a new test, it has to be proven that it would make a difference to patient outcome…
In 2015, the National Institute for Health and Care Excellence in the UK (NICE) produced diagnostics guidance7 on using procalcitonin testing for diagnosis and monitoring sepsis in adults and children. Chapter 5: Outcomes gives a good summary of the evidence that was available to them at the time, in order to make their recommendations. They looked at 8 studies from an intensive care setting which investigated the use of procalcitonin to cease antibiotic therapy, and 10 studies from an emergency department setting where the focus was on using procalcitonin to guide a decision whether or not to start antibiotics. The papers they included compared standard practice with standard practice plus procalcitonin testing. The outcomes were around antibiotic exposure, resource use (hospital admission/length of stay/cost), and adverse events.
They found that, in an intensive care setting, procalcitonin testing significantly reduced duration of antibiotic therapy in all studies (-3.19 days, 95% confidence interval -5.44 to -0.95), significantly reduced length of stay in 2 studies with no difference in the others, and there was no difference in adverse outcomes reported. In an emergency department setting, procalcitonin testing significantly reduced antibiotic use in adults (relative risk 0.77, 95%CI 0.68 to 0.87), reduced antibiotic course duration (though not statistically significant except in one paediatric study), and no difference in adverse outcomes.
Their cost analysis found that procalcitonin testing with standard clinical practice remained cost effective compared with standard clinical practice alone. An economic report published in 2010 by the NHS Centre for Evidence-based Purchasing also found procalcitonin testing to be cost effective, though this report looked solely into treating patients with lower respiratory tract infections.
On considering the above, NICE found that procalcitonin tests show promise, but that there was insufficient evidence to recommend their routine use within the NHS. They did acknowledge that there are departments in hospitals around the country that are using it at present and recommended further research take place.
What’s happened since then?
Well, sadly not a lot. There have been a few more studies looking into procalcitonin, like this one8 of 318 infants under the age of three months presenting to the emergency department with pyrexia of unknown origin. The team from Madrid found that procalcitonin was a better diagnostic marker than others such as CRP, white cell or neutrophil count, but only identified 70% of patients with invasive bacterial infection. Certainly not enough to dispense with the antibiotics just yet. A systematic review published in 2016 of the literature surrounding procalcitonin use in guiding antibiotic prescribing in the ED came to the same conclusion as the NICE report9. Earlier this year, a team led by Jørgen Vestbo at the University of Manchester and the North West Lung Centre carried out a meta-analysis of the literature around procalcitonin use in treating patients with COPD10. They found evidence that procalcitonin-based protocols could decrease antibiotic prescription and antibiotic exposure without affecting clinical outcomes, however they acknowledged the quality of that evidence is low to moderate due to small study population. They stressed the need for appropriately designed and powered randomised controlled trials in this population, and really that can be extrapolated to a call for the same in any group. We’re not much further along than we were in 1983.
So what can we draw from all of this? Well, procalcitonin has been extensively studied, but sadly each time it seems in a different population with a different question in mind, be it adults with pneumonia or children with all-cause sepsis. There are studies on procalcitonin use in identifying bacterial infection in patients with burns11 or cancer12. Even in those studies which are similar, often quite different conclusions have been drawn about the usefulness of the test. With so many varied studies, it’s difficult to come to an agreement one way or the other, but it seems at the moment that it’s not the fix we’re looking for to absolutely determine whether our patients need antibiotics or not. Maybe in certain circumstances we can add it to our arsenal of tests, but currently it doesn’t feel like it would add too much to my practice in the ED. We certainly need some good RCTs to help us along.
It would be interesting though to hear views from around the world on this 34 year old test. Do you use procalcitonin in your emergency department/intensive care, and who for?
Before you go please don’t forget to…
- Subscribe to the blog (look top right for the link)
- Subscribe to our PODCAST on iTunes
- Follow us on twitter @stemlyns
- PLEASE Like us on Facebook
- Find out more about the St.Emlyn’s team
10 thoughts on “Are you PROcalcitonin? St.Emlyn’s goes bug-hunting.”
Pingback: Are you PROcalcitonin? St.Emlyn’s goes bug-hunting. – Global Intensive Care
From an ID doc’s perspective my take is that if you are somebody who does a comprehensive clinical assessment and has a clear justifiable rationale for using antibiotics in each case, it probably doesn’t influence decisions too much (the ‘clinical impression overrides the test’ element of the recommendation is the most weighted). It’s a bit like why CRP in ID circles also goes by CRaP as a diagnostic test. Your diagnosis is essentially unchanged whether or not you are blinded to the CRP result. Try it…
However, there are many people who think antibiotics are there for ‘just in case’ ‘what’s the harm’ scenarios, where their rationale for initiating therapy does not have watertight justification or who are reluctant to stop in the absence of evidence of improvement or results to support bacterial infection. So for the purpose of giving the latter group a crutch to make them feel less uncertain about withholding antibiotics, it’s probably useful. And certainly less labour intensive than education drives. When we consider that antibiotics are the only class of medicines we have with transmissible loss of efficacy over time, anything that rationalises (but not rations) their use is welcome.
Great post Chris, really clear summary of the role of procalcitonin
As you said, it’s been around for a long time and hasn’t become routine use. There are lots of studies and lots of people can argue both sides very very well. It’s expensive relatively and isn’t cost neutral which is best you can really prove or achieve. It adds evidence to the range we have but clinical experience overrides. The units that use them and the units that don’t there aren’t really any obvious difference in their outcomes or their antibiotic stewardship. The algorhythm can be easy but also complicated…
It’s difficult to get passed as a business case at the moment and I’m not sure it would add or give benefit to the critical care population
Also not for use in ed as it’s lag time is too long… you need base line and day one data to know really (if low on day one doesn’t rule out bacterial infection)
It’s ben around for ages and it will be and unlikely to be come common place I think
The units that don’t use CRP in critical care have the same if not better infection and antibiotic use age
There are loads of innovation around identification of pathogens coming out all the time using neeer techniques… probably better to invest time and resources into this if after so long no clear advantage has been shown
My other bug bare is the thought that tazocin is a stringer anribuotuc than co amoxiclav or amoxicillin… it isn’t and failure of antibiotic therapy is not after 1 or 2 doses!!!
If we are going to use any biomarker as a decision tool, it would seem that the first important step is to obtain a pre-test probability based on clinical risk assessment.
Although there are a few good isolated examples of this e.g. Wells score, I don’t we put enough emphasis on this in general when teaching/practising clinical medicine.
We know has a general rule that even tests with high accuracy are best employed (and most cost-effective) when there is intermediate probability of disease. If there is already low suspicion, the chances of creating false positives is increased (resulting in over-treatment). If it is high, then the chances of false negatives is increased (resulting in under-treatment). Ordering CRaP indiscriminately results in generating a lot of CRaP.
The question isn’t so much whether it the test is 100% accurate but whether it improves the predictive values above what our current tools allow us (including the option of watchful waiting). And carefully choosing the populations we are testing will significantly impact the cost-benefit of the test.
Absolutely a key point Derek, and yet there are many pressures in the system to test early and to test without forethought. This often leads to overdiagnosis which is increasingly a problem.
Four hour targets being the most vexatious cause. But some of our junior docs are led to believe they can apply the same investigative strategies to the ED population compared to the sicker hospital population. They forget that a selection process (including but not limited to clinical assessment) has already occurred prior to the patient hitting the ward. This leads to bracket creep with some of these practices filtering even down to general practice (where the majority of them end up practising).
Pingback: LITFL Review 279 • LITFL • Life in the Fast Lane Medical Blog
I am not that pro about procalcitonin. It is not the magic bullet to help us take all important decisions regarding antibiotic stewardship.
It can at best be used as an adjuvent to guide therapy and predict prognosis.
Check out my resource on PCT at.