Are you ready for an acute inflammatory mediator for sepsis that is consistent and works just as we’d expect it to? Well, it’s here: procalcitonin. Or is it? It can certainly claim that it works just as well as former acute phase reactants and inflammatory mediators – including erythrocyte sedimentation rate (ESR), the c-reactive protein (CRP), interleuken and the white blood cell count (WBC). But these predecessors all suffer from a similar, critical problem - they are somewhat sensitive but not very specific. After review of the literature, I don’t see procalcitonin producing much different results.
Let’s get right to business. Many would have you believe that procalcitonin is something different from the pack. After 15 years of being touted as the most promising inflammatory mediator, maybe it’s not as promising as once thought. In the current context, it is said to possess the ability to differentiate those with systemic inflammatory response syndrome (SIRS) without infection from those who are infected. In other words, this test is supposed to tell you who’s septic and who isn’t.
In 1498, Niccolo Michiavelli, an Italian historian, diplomat and political theorist, stated, “Hectic fever (sepsis) at its inception is difficult to recognize but easy to treat. Left untended, it becomes easy to recognize but difficult to treat.” His assessment holds true today and is the cornerstone of the issue with diagnosing sepsis. In my opinion, inflammatory mediators follow this same pattern, sensitive in severe sepsis and not so in early infection.
Good for determining septic shock? That’s a lot like saying a thermometer is a great tool for determining if your house is on fire. Sure, the temperature is really high, but the flames and smoke pretty much clued you into the inferno already. Procalcitonin and septic shock is a similar situation. We don’t need a marker for sepsis (inflammatory mediator) when making the diagnosis, once the patient is in septic shock, is so easy that the Geico “caveman” could do it. I’m certain the ESR CRP and a host of other inflammatory mediators would be elevated also; PCT adds nothing to this diagnostic picture. In contrast, when the infected patient is either not septic or in early sepsis, the test cannot predict who will develop sepsis and is more likely to remain below the reference range until enough inflammatory response has occurred to generate a PCT above the reference range. Thus, when we need the help, detecting early sepsis or predicting who will develop sepsis, the test is less helpful, and when the test is likely to be positive, we don’t need the help, as the diagnosis is clinically available.
Different phases of disease and different host responses make the business of applying PCT values to an individual patient nearly impossible, and thus, problematic to universally adopt. For instance, for those who may not mount a strong immune response, such as the elderly, less inflammatory response may result in less accuracy of acute phase reactants. Authors have reported that procalcitonin is nearly non-detectable in healthy hosts and elevates rapidly to its maximum concentrations between 6-12 hours of infection (Medicina (Kaunas). 2006;42(1):69-78.). However, we’re not comparing healthy individuals to sick individuals. Mitaka (Clin Chim Acta. 2005 Jan;351(1-2):17-29.) and many others have reported that PCT increases with the severity of sepsis and it does so more predictably than CRP. I will concede that point. First, we need to know, of the sick, who’s septic and who isn’t. As sepsis worsens, our diagnostic burden is lessened. Second, we need to by able to apply aggregate data to a single patient. The fact is that we are relying on a host response, which is variable, to result in a predictable mediator response. I believe this is why, after 15 years, the best in bread of mediators still may not be able to answer our most important questions.
In review of the literature, I will concede three points: Procalcitonin elevates as infections worsen, it is more predictable than other inflammatory mediators and it may guide antibiotic use in septic patients, particularly in the ICU and after around 3 days of treatment. I cannot except that PCT will tell us which patients are infected, who will go on to develop severe sepsis or that this market can differentiate a viral, from a bacterial source of infection.
Reduction of antibiotic use is an important benefit of PCT that must be highlighted. However, this benefit has no value for the emergency department patient. In a randomized, controlled trial, antibiotics were discontinued in critically ill patients when the PCT decreased by 90% from the baseline. The median duration of antibiotic administration was 3.5 days less for the PCT group compared to the empiric treatment group. However, the antibiotics were not discontinued prior to 3 days of therapy (Am J Respir Crit Care Med. 2008 Mar 1;177(5):498-505. Epub 2007 Dec 20.).
Although claims have been made that PCT levels can differentiate the actual etiology of an infectious source, this just doesn’t seem reasonable that a single test could do so when any infectious etiology may result in a similar inflammatory response and cascade being triggered. One study reporting on PCT and the severity of pneumonia (Journal of Infection Volume 52, Issue 3, March 2006, Pages 169-177.) reported that PCT had elevated levels in pneumococcal infections, low levels in Legionella and undetectable levels in pulmonary tuberculosis and pneumocystis pneumonia.
In detecting sepsis many of the studies, reporting positive data for any of these inflammatory mediators are observational studies, comparing which mediator performed best in septic patients, predicting septic shock or even predicting mortality. Universally, PCT wins those battles. Many studies have shown that PCT outperforms CRP; but we don’t need to know which marginal tool works better than another (Anaesth Intensive Care. 2002 Dec;30(6):747-54.) To that end, 64 consecutive pediatric ICU patients with SIRS were tested with PCT and CRP to determine bacterial from non-bacterial etiologies. PCT was better. However, both were reported to only have “moderate accuracy” (Pediatr Crit Care Med. 2008 Jul;9(4):407-13.). That is the wrong question to be asking. “Which performs better in sepsis” doesn’t help us determine which, if any, can differentiate sepsis from non-sepsis or infected from non-infected. Those are the real questions, and I doubt any sepsis biomarker will ever be able to do that on an individual patient basis. Tang, et al. published a meta-analysis in 2007 stating that PCT cannot reliably differentiate between infectious and non-infectious causes of SIRS (Lancet Infect Dis. 2007;7:210-217.). Mueller, et al. wrote the following in response: “A gold standard to differentiate infectious from non-infectious causes in patients with SIRS [systemic inflammatory response syndrome] does not exist, and therefore all observational studies are prone to a potential bias.” I hope they didn’t spend much time writing that rebuttal. I find it less than persuasive. So, because no gold standard exists, you have to give us a pass on the limitations of our data? Sorry, no such pass given, Doctor. Please prove that your test does what you report it will do before expecting us to drop our guard and apply this to the masses.
Several other recent articles cast doubt on PCT’s credibility for mainstream use. “The prognostic utility of initial procalcitonin measurement in respiratory tract infections is suboptimal.” Low presenting PCT levels should be outweighed by higher absolute risk for adverse outcomes (Swiss Med Wkly. 2009 Jun 13;139(23-24):318-26.).
Although many studies have shown a marked reduction in antibiotic use without a significant difference in mortality, one article reported a higher than acceptable variation in mortality ranging from 7.5%-10% in some studies ( Int J Antimicrob Agents. 2010 Dec;36 Suppl 2:S17-21. Epub 2010 Dec 3.).
Lanoix, et al. reported that in patients with acute SLE flares, PCT levels were within the normal range in both infected and non infected patients, with flares and without flares (Lupus. 2011 Feb;20(2):125-30. Epub 2010 Oct 11.).
The VIDAS BRAHMS PCT test, bioMierieux, seems to be the manufacturer on the forefront pushing this innovation. Although useful technology should have its economic rewards, we need to make certain that such rewards don’t result in undo influence of providers and institutions if the data doesn’t support the assertions. It is interesting to note that the only two articles appearing on the companies website are:
1. Effect of procalcitonin-guided treatment on antibiotic use and outcome in lower respiratory tract infections: cluster-randomized, single-blinded intervention trial. Christ-Crain M. et coll. Lancet. 2004 ;363:600-7; and
2. Morin N., Renard N., Gaude H., Hainque B., Bernard M. Evaluation Of The Analytical Performance Of Vidas® B·R·A·H·M·S PCT, Poster presented during the CORATA, Oct 24-27 2006, Bruges.
If the data overwhelmingly supports the use of this product and technology, why not point us to the resources that support your product’s use? What is also important is to disclose are any pertinent conflicts of interest. Although a physician with impeccable credentials, in January 2010, Dr. Amin participated in a webinar titled, “The Early Identification and Management of Septicemia in the Acute Care Setting.
His disclosures included, “Consultant and speaker for BioMérieux Speaker for Actelion,” and this must be factored into the weight of his opinions.
A higher burden of proof should be needed to adopt a new test or new application of an old test. Based on this premise alone, this debate is over; that burden of proof has not been met. You can often times find data on either side of an argument. PCT and sepsis is a testimonial to that fact. Data on both sides of an issue equals a lack of consensus.
Although it is reasonable to say that PCT is probably the most promising of sepsis biomarkers, and may, in fact, reduce antibiotic usage, where emergency physicians need help the most, PCT doesn’t offer much assistance. The day may come where PCT proves its worth, but that will likely be in combination with clinical judgment and perhaps, an evidence-based decision rule.