Are There Pros to Procalcitonin?

The Gist:  Procalcitonin (PCT) is gaining popularity in identification of serious bacterial infections and in attempts to curtail antibiotic usage; however, it has many limitations and has not been proven to conclusively and consistently alter patient oriented outcomes.  An elevated PCT may indicate infection, but this goes out the window in trauma or surgery.  The data seem to support using PCT over other biomarkers to help inform length of antibiotic use, especially in respiratory infections (i.e. if it falls to <0.5 mcg/L then it's likely safe to cease antibiotics) but any patient benefit derived from this is unclear.  This benefit is derived after the decision to initiate antibiotics has been made.  

As part of my clerkship, I recently attended a lecture on the use of PCT in the ED and ICU for sepsis identification and management.  The lecture was overwhelmingly promising - a better biomarker to aid in an attempt to sort out bacterial infections and stratify sick from less-sick patients in clinically murky situation?  Sounds awesome, but the lecture was industry sponsored and FOAM has taught me to ask a few questions about any diagnostic method I employ.  What does the test actually indicate (if positive or negative)?  How good is it, really?  How will it change my management of this patient?  Is it costly?  Timely?  Thus, I decided to investigate the test for myself.  Naturally, I went to my trusty FOAM sources to aid in this endeavor.

• I started with PulmCCM.org, where I recalled reading several things on procalcitonin.  They have an fantastic review on use of procalcitonin and pneumonia, which also answers some of the more general questions about procalcitonin.  They also have a summary of some recent literature reviews. 
• Then however, it was time to go to the pool, so I found a free talk by Dr. Talan on procalcitonin in the ED.  
• In typical FOAM fashion, two days later, a brand new podcast on PCT (episode 194) by the Society of Critical Care Medicine started playing while I was at the gym.  It's a good review of some recent literature and some projections for the future of PCT.

Basics of Procalcitonin

• More specific for bacterial etiology compared with other markers such as WBC and c-reactive protein (CRP) due to interferon-gamma from host response to virus (1)
• Short half-life so theoretically easily trended and reflects patient's current situation (1)
• Not as altered by inflammatory states, glucocorticoids, or other known drugs as WBC and CRP; however, it does rise immediately after major trauma and surgery in the absence of bacterial pathogens (1)
• Costs ~$25 per test in the US (charge is ~$38)

Below is essentially a run-down of the primary sources referenced in the above reviews (it's covered by the FOAM sources above, but I'm trying to get better at going through the primary literature).  Note:  most of these studies looked at PCT use in patients with respiratory symptoms, many studies are funded by the manufacturer, and some of them use various cut-off points for interventions.  I've also left off some of the old studies that used the older assay not currently widely used in the U.S.

Identification of Serious Bacterial Infections (SBI) - not ready for prime time.

• Sudhir et al found that PCT was "positive" (>0.5mcg/L) in 94% of their 100 patient septic cohort (older assay)
• 2010 pediatric study randomized febrile kids to either have a PCT value available to the provider or not.  Used a cut off of 0.5 mcg/dL.  Sensitivity of PCT for SBI 77%, Specificity 64%
• June 2011 study by Bafadhel et al evaluated PCT and CRP values in Community Acquired Pneumonia (CAP) versus COPD/asthma.  For identifying pneumonia, CRP >48 mg/L had a sensitivity of   91% (95% CI, 80%-97%), specificity 93% .  A PCT of 0.8 mcg/L had a sensitivity of 89% (95% CI, 78%-95%) and specificity of 78% (95% CI, 72%-82%). 
• Meynaar et al evalued PCT in 76 patients (cut offs of 2 mcg/L and 10 mcg/L) and have a nice table of the Se, Sp, and OR compared with other biomarkers
• This study by El-Solh et al in Critical Care Medicine 2011 did not differentiate between aspiration pneumonitis and bacterial aspiration pneumonia.
• Muller et al published an analysis of the ProHosp data and found that patients with PCT levels <0.25 mcg/L were extremely unlikely to have positive blood cultures (<1%) 
• Reynolds et al found that shock elevates PCT levels regardless of bacterial infection in a study of medical ill and post-operative patients.  Higher PCT levels were assosciated with infection, but this data support the notion that general inflammation does affect PCT levels regardless of the presence of bacterial pathogens. 

Clinically meaningful outcomes for patients
Morbidity/Mortality - probably no difference in mortality using PCT guided algorithms, questionable change in length of stay

• In the ProHOSP RCT by Schuetz et al in 2009, utilization of PCT did not reduce the incidence of adverse effects (primary outcome of the study), although the authors conclude that the incidence was "similar." 
• Another 2011 systematic review by Schuetz et al (who published many of the PCT trials) showed no mortality difference in patients with PCT guided therapy versus standard care groups despite reduced antibiotic usage in the PCT guided cohort

Length of Stay (LOS) - unlikely that PCT has any effect. Would you discharge someone based solely on a lab value?  I'm pretty sure everyone uses the clinical picture in this setting.

• In Agarwal and Schwartz's systematic review, only 2 studies demonstrated decreased LOS.
• Jensen et al's 2011 study in Critical Care Medicine investigate the effect of using an algorithm based on PCT values to escalate antimicrobial therapy (escalated at 1 mcg/L, a higher cut-off level than the 0.5 mcg/L used in most other studies).  The study, reviewed here by EM Lit of Note, demonstrates that there's actually an increase antibiotic usage, LOS, and ventilator dependent days.  

Antibiotic Usage - studies pretty consistently demonstrate that trending  PCT can reduce the length of antibiotic usage in respiratory infections but should not be used to initiate antibiotics.

• June 2011 study by Bafadhel et al  Derivation of PCT values to guide therapy: PCT value of > 0.25 ng/mL would have incredibly reduced antibiotic usage in patients with asthma or COPD exacerbations while slightly reducing the antibiotic usage in the pneumonia cohort.  Counterpoint:  A CRP value of >48 mg/L also resulted in very similar reductions
• Asthma - from 57% to 4% (reduction in antibiotic usage of 93%; 95% CI, 88%-98%)
• COPD - from 76% to 7% in patients with exacerbation of COPD (reduction of 91%; 95% CI, 87%-95%)
• Pneumonia - from 100% to 73% (reduction of 27%; 95% CI 17%-40%).
• Antibiotic duration in PCT group 5.7 days vs 8.7 days in standard care group in the ProHosp trial 
• A 2011 systematic review by Agarwal and Schwartz (including only RCTs) concluded that use of PCT to inform antibiotic length decreases usage of antimicrobials.
• The primary outcome of the  2010 pediatric study was antibiotic usage, which was the same in the group who had PCT available and those that did not.  If antibiotics were given to all kids with a PCT >0.5 mcg/L, more antibiotics would have been utilized (24% 95% CI, 15-33), not less.  Additionally, there's an argument that an even lower cutoff value should be employed (0.25 mcg/L has been proposed)
• Schuetz et al's systematic review  concluded that PCT reduced antibiotic duration and prescriptions in the outpatient, ED, and ICU settings for patients with respiratory infections (note: included more trials than the Agarwal and Schwartz trial). 
• Not all studies demonstrating a statistically significant reduction in the exposure to antibiotics, including the 2009 Schuetz RCT, used this as a primary endpoint in the study.  In this cohort for lower respiratory tract infections, the mean length of antibiotic exposure was 5.7 days in the PCT group compared with 8.7 in the standard care group. 
• Limiting initiation of antibiotic therapy is probably not effective at a cut-off of 0.5mcg/L (so use your clinical judgment) (2)

Risk Stratification 

• EPs are good at determining who is extremely ill and those who aren't that sick but in situations where the patient can seem to go either way, lab tests may help. Harbarth et al found that patients on the SIRS/sepsis spectrum had poorer prognoses if their PCT levels didn't fall (3).
•  High risk:  PCT levels >2 mcg/L.  
• Low risk: < 0.25 mcg/L  
• The cut-off of 0.5mcg/L seems to be the most referenced

Economics - the test is presently more expensive than the vague counterparts, the WBC count and CRP.  However, some argue that there are cost savings in LOS reduction, antibiotic usage, blood cultures (using PCT as a surrogate - see Muller et al Table 5), and reduced ICU days.  Presently, most of these arguments are projection of data based on derivation and internal, retrospective validation of certain PCT cut off and theoretical algorithms.  

PCT is gaining steam in the US.  The test llooks better than WBC and CRP so far, but studies are beginning to show that the test has some confounding issues.  Exercise caution when ordering PCT and recognize the soft lines drawn for the "cut-off" levels of the test.

References:
1.  Jin M et al Procalcitonin: Uses in the Clinical Laboratory for the Diagnosis of Sepsis Lab Medicine 2010 41, 173-177 
2.  Agarwal and Schwartz.  Procalcitonin to Guide Duration of Antimicrobial Therapy in Intensive Care Units: A Systematic Review Clin Infect Dis. (2011) 53 (4):379-387.
3. Harbarth et al.  Diagnostic Value of Procalcitonin, Interleukin-6, and Interleukin-8 in Critically Ill Patients Admitted with Suspected Sepsis Am. J. Respir. Crit. Care Med.August 1, 2001 vol. 164no. 3 396-402

P.S. - This Not-So-Sick Looking Patient May Be Septic

The Gist:  Exercise caution in febrile asplenic patients because these folks are at an increased risk of post-splenectomy sepsis (PSS), particular encapsulated organisms and red blood cell (RBC)-based parasites.  Sepsis may present differently in these patients - maintain a high index of suspicion.  For a good case-based podcast review, check out the Splenectomy-Sepsis lecture (Episode 26) from UC-Irvine.  

This isn't crazily, common.  Do I really need to care?
Somewhere between a hematology lecture and my surgical rotation, I became enraptured with the spleen.  A typically unassuming organ, the spleen bridges innate immunity with adaptive immunity, houses 30% of RBCs, and cleanses the dregs of our cells.  

• It turns out that despite the fact that patients are benefited by splenectomy, there are very important, preventable complications from our patients.  PSS is likely the most feared of these complications and is decreasing in frequency due to the utilization of prophylactic vaccines.  However, these patients become sick very quickly and may present with an atypical picture.  Furthermore, this may occur in young, otherwise healthy looking folks who can compensate (aka not your typical sepsis picture). 
• Variable numbers of  asplenic patients aren't aware that they're at increased risk for infection - including up to 84% in a 2000 publication (2) and 50% in a 2008 paper (3).  Thus, the patient may not disclose that they've had a splenectomy.  In these patients, history can be crucial.
• Check for splenectomy scar
• Ask about travel, tick exposure, vaccination status, sick contacts.

Clues that a patient may be asplenic:

• Hemoglobinopathy (sickle cell anemia, thalassemia) 
• Left sided abdominal surgical scar due to:
• Hemolytic anemia (e.g. hereditary spherocytosis)
• Trauma. Note:  S/P blunt trauma - sepsis may be less common due to the presence of accessory spleens
• Assorted other causes: thrombocytopenia, hypersplenism, malignancy

So, how do they present?

• PSS is most common in the first few post-splenectomy years but can occur for decades after splenectomy.
• Fever and rigors are the most common signs.  
• Headache and gastrointestinal complaints (diarrhea) can be manifestations of severe pneumooccal infections and may be mistaken for gastroenteritis
• Disseminated Intravascular Coagulation

Suspect Organisms

• Encapsulated bacteria (due to the production of IgM memory B cells, specially located within the spleen):   Streptococcus pneumoniae, Haemophilus influenza, and Neisseria meningitidis.  Also,  Capnocytophaga canimorsus from dog bites.
• Parasites:  
• Malaria (Think travel to endemic areas like S.Asia, sub-Saharan Africa, etc)
• Babesiosis - tick borne illness, especially in NY/CT/MA and s/p RBC transfusion 
• Ehrlichiosis - tick borne illness in the Southeast/mid-Atlantic US (1)
• Virus:  Cytomegalovirus

Patients with PSS deteriorate quickly, what should I do?

• If they're septic, naturally follow the sepsis protocol.
• Patient looks sick with a questionable/unknown vaccine history?  Give empiric antibiotics.
• Ceftriaxone Adults: 2 g IV q12 to 24 hours, Children: 50 mg/kg IV q12 hours (1)
• Vancomycin if S.pneumoniae penicillin resistance suspected:  Adults: 1 to 1.5 g IV q12 hours, Children: 30 mg/kg IV q12 hours (1)
• Thinking babesiosis?
• IV clindamycin and oral quinine or IV atovaquone and IV azithromycin
• Have a high index of suspicion for meningitis, so have a low threshold to do an LP if warranted.

Pretty sure my patient really does just have the flu/gastroenteritis/etc, what do I do now?

• It's recommended that asplenic patients receive prophylactic antibiotics at the onset of febrile illnesses, although this recommendation is not evidenced based.  For example, British guidelines recommend  Amoxicillin prophylaxis through adulthood.  
• Those with adequate primary care physician coverage receive vaccines against pneumococcus, meningococcus, and haemophilus but even these are not fool proof (interestingly, over-vaccination with pneumococcal vaccine actually has a counterproductive effect on the immune system).
• Educate. Even if they have something minor, let them know when to return to the doctor.

References:

1.  Pasternack, M.  Clinical features and management of sepsis in the asplenic patient.  

2.  Di Sabatino A; Carsetti R; Corazza, G. Post-splenectomy and hyposplenic states. The Lancet378. 9785 (Jul 2-Jul 8, 2011): 86-97.

3.  Wilkes, A, Wills, V, Smith S.  Patient knowledge of the risk of post-splenectomy sepsis.  ANZ Journal of Surgery. Volume 78, Issue 10, pages 867-870, October 2008.

Surviving Sepsis -Leveling the Playing Field via the Web

The Gist:  Over the past decade, the Surviving Sepsis Campaign and Early Goal-Directed Therapy have altered the perception and management of sepsis  These guidelines and other crucial ED interventions change frequently, as evidenced by upcoming changes to the sepsis guidelines.  Yet, differences remain between academic and community settings in the implementation of these protocols.  With the help of FOAM there is hope in closing the disparity!

I've completed nearly all of my third year clerkships at two community hospitals that serve a rather large, multi-state rural catchment area (the primary hospital's ED has >57,000 visits/year).  The clinical experience is incredible and I train with outstanding, bright physicians.  As I follow podcasts, blogs, and tweets, however, I've discovered that the academic accepted "standard of care" or "best practice" often differs from community practice.  I was dismayed to find that our EDs don't use waveform capnography in intubations/procedural sedation, use ultrasound outside of FAST or vascular access, and only initiated post-arrest hypothermia within the past few months.  (Note:  I dislike referring to community and academic EDs in such a homogeneous fashion, as I'm aware there is quite a spectrum of practice, but these terms align with current discourse and seemed suitable.)

The paper that got me thinking:  Early Goal-directed Therapy (EGDT) for Severe Sepsis/Septic Shock:  Which Components of Treatment are More Difficult to Implement in a Community-based Emergency Department?

• Assesses the results after implementation of an EGDT protocol for severe sepsis and septic shock in a community hospital (>65,000 visits/year) from 2008-2009.

Result Highlights

• 66/85 patients received antibiotics in first hour
• 58/85 patients received a 2L fluid bolus in first hour (68%)
• These patients were more likely to survive to hospital discharge
• 79% (n=50) of patients needing vasopressors received these drugs.
• Central line in 55/85 patients (65%, 95% CI 54–74%)
• CVP recorded in the ED 23/85 patients (27%, 95% CI 18–36%).  
• ScvO2 (central venous oxygen saturation) measured in 13/85 patients

But what do these results really point to?

• This paper found that the most invasive components (A-lines, central lines) of their resuscitation bundle were followed less frequently.  The authors concluded that this may have partially resulted from providers/staff feeling uncomfortable with these invasive procedures.  This rationalization seems odd as this community based ED also houses an EM residency program, but this is one of the most frequently cited challenges to sepsis protocol (solution = procedure hungry medical students?).    
• Furthermore, the CVP and ScvO2 were rarely recorded, even in the presence of a central line.  The CVP has been pretty much discarded, but the ScvO2?  Did these providers get the information they needed to direct therapy in other ways?  If so, it's not documented in this article.  It's also unlikely, given that the Jones article in JAMA on lactate clearance wasn't published until 2010.
• The fluid resuscitation goal of at least 2L of fluids in the first hour was only met in 68% of the patients.  This goal was pre-determined and didn't focus on patient response to boluses.  This is one of the seemingly easiest measures to implement and, in this particular patient population, was associated with a mortality benefit.  

Is there a solution?

By following EM literature and updates in an asynchronous and self-directed method, I see great alternatives/solutions to some of these issues and barriers.  For example, EMCrit's CME option could act as an incentive/inspiration for community doctors to implement cutting edge EM locally...while satisfying required CME (see real-life example on PHARM podcast below).

• Really understand the importance of sepsis and champion excellent, early sepsis care. Parts I, II, and III from EMCrit (with Dr. Rivers himself). 
•  I admit that I kept these videos on my iPhone/computer for quite some time before I ran out of my favorite podcasts and needed something to get me through a workout.  Sepsis just didn't seem as sexy as all of the airway and trauma talk but this definitely changed my perspective and the videos are quite good...I think I'm probably not alone in my initial attitude (especially if you look at the compliance rate with the severe sepsis protocol in the above study).  To entice other medical students there's a Septris game from Stanford.
• Assessing Fluid Responsiveness with Ultrasound and Passive Leg Raise. Courtesy of the amazing Ultrasound Podcast parts 1 and 2.  Only have a minute and a smart phone? 

                                         

• Lactate Clearance!  The use of non-invasive sepsis protocols that use lactate clearance instead of ScvO2 are somewhat controversial but, in any case, are better than not measuring either (referencing above study), especially in your less sick severe sepsis patients.  Dr. Scott Weingart's EMCrit site has non-invasive sepsis protocols and answers to popular questions about lactate.  I first learned about this on the excellent EMCrit Podcast 22. 
• What about pressors?  Free Emergency Medicine Talks comes through again with Dr. Evie Marcolini's lecture on pressor choice.  She also has a talk on antibiotic choice in sepsis.  
• Want to put it all together?  Dr. Weingart has an entire Severe Sepsis Protocol available.  

Beyond sepsis...the podcast that kept me thinking:

There is hope for clinician driven change in the community-based setting.

• I found inspiration in a recent episode #16 of the PHARM podcast on delayed sequence intubation (DSI) where Dr. Rob Bryant shares his experience implementing DSI in the community setting.  He shared his DSI protocols with others via PHARM and EMCrit.  It is truly amazing to see such dissemination of knowledge and experience, flattening the playing ground for clinicians.  

References:

O'Neill R, Morales J, Jule M.  Early Goal-directed Therapy (EGDT) for Severe Sepsis/Septic Shock:  Which Components of Treatment are More Difficult to Implement in a Community-based Emergency Department?  Journal of Emergency Medicine.  Volume 42, Issue 5, May 2012, Pages 503–510

Urine Albumin:Creatinine Ratio - The Next Big Thing or an Excuse for Substituted Judgment?

The Gist:  Urine Albumin: Creatinine Ratio (ACR) is higher in the sicker septic patients compared with the less sick septic patients and may be able to predict which patients will take a turn for the worse.  This test probably won't add much to clinical care or goal directed therapy, at least in the ED at this time.  

Sepsis is kind of a big deal.  The "Surviving Sepsis" campaign championed by Dr. Rivers seems to dominate EM literature and we continuously seek tests that allow us to identify, stratify, and treat septic patients.  According to a study in American Emergency Medicine last month, patients presenting to the ED with sepsis do at least as well as those patients who are directly admitted to the hospital.  Good job.  However, we can always do better.  

As in other high-risk issues like chest pain, myocardial infarction, and pulmonary embolism, we look for a test that will allow us to pat ourselves on the back for making a decision we've already really made using clinical gestalt.  In pulmonary embolism, this may be a d-dimer or CT scan in a PERC negative/low-risk Wells patient.  Lactate, central venous pressure (CVP), and inferior vena cava (IVC) diameter are some of the parameters currently utilized to help stratify septic patients.  The urine albumin:creatinine ratio seeks to do the same.  This test is based on the fact that the kidneys tend to spill albumin during inflammatory states, a result of endothelial cell injury.  Thus, the sicker patients should spill more albumin in their urine, perhaps before they deteriorate into a state of tissue hypoxia.    

This study was published in Academic Emergency Medicine in March 2012 and is predicated on the theory that the ACR could predict which sepsis patients have more severe outcomes.  

Methods

• Convenience sample of patients meeting criteria for sepsis presenting to one of two EDs
• Pilot study - prospectively enrolled if the patient had a urinalysis done
• Patients could not have signs of tissue hypoxia or shock at the time of presentation
• Primary outcome was disposition from ED (ICU/step down, discharged home, admit to floor)

Results

• Patients admitted to more intensive medical care settings had higher levels of ACRs compared with those patients who were sent home.
• This didn't hold true for patients with a genitourinary source
• Better prognostic indicator than serum lactate levels (which were not prognostic in this cohort).
• So, patients who already ended up in the ICU or on the floor due to clinical presentation/labs had higher ACR
• Patients had a solid mean arterial pressure (MAP), with an average MAP of 89.5. 
• This patient population wasn't comprised of the super sick, although 51 of the 92 patients qualified for the title "severe sepsis."
•  Perhaps the physicians were more cognizant of the disposition they gave to their patients, indicative of a Hawthorne effect.

In the future, this test may play a role in allowing physicians to feel better about discharging a patient home or  convince an intensivist to accept a patient who might have otherwise decent labs or vitals.  It may increase our sensitivity in detecting septic patients that will deteriorate.  My best guess, however, is that it really only tells us what we already know in our clinical assessment of a patient.   

References:

Drumheller, B. C., McGrath, M., Matsuura, A. C. and Gaieski, D. F. (2012), Point-of-care Urine Albumin:Creatinine Ratio Is Associated With Outcome in Emergency Department Patients With Sepsis: A Pilot Study. Academic Emergency Medicine, 19: 259–264. doi: 10.1111/j.1553-2712.2011.01266.x