Jeff: So as well all know, sepsis is bread and butter emergency medicine, but, what is sepsis? It seems that every month or so we have a new guideline, bundle, definition, or whatever… I think it’s best to start with the basics - At its core, sepsis is a dysregulated response to infection that can be life-threatening.
Nachi: Right and it’s the combined inflammatory with immunosuppressive features of sepsis that lead to the devastating organ dysfunction and even death. Optimal management of septic patients has been a source of intense research, stemming from the landmark study by Rivers in 2001. Jeremy, can you give us a little historical context there?
Jeremy: Rivers was a real pioneer. He found a 16% mortality reduction with randomization to an early aggressive care bundle. Amazing work. That being said, many components of that bundle have since been disregarded. For example, Manny Rivers would measure CVP in all of his patients, something we rarely do.
Nachi: Not to cut you off and steal your thunder there, but we’ll get to the most recent updates in management shortly. Let’s first talk definitions and terminology, and specifically, diagnosis, which is definitely a big elephant in the room. As Jeff mentioned a few minutes ago, diagnostic criteria have undergone so so so many changes.
Jeff: Yes it has! 1991 marked the first standardized definition. Then in 2001, sepsis-2 was introduced. In 2014, the Society of Critical Care Medicine and the European Society of Intensive Care Medicine started a task force, and by 2016, updated definitions were out again! Sepsis-3!! A lot of this came after the realization that SIRS was just too broad and was overly sensitive and non-specific. Jeremy, why don’t you take us through Sepsis 3.
Jeremy: So just to back up a little and frame this:
Here’s the fundamental problem: As we like to say, “there’s no troponin for sepsis.” And if you look at our patients, we tend not to miss the hypotensive, tachycardic, febrile patient. We know they’re septic. But how do we find the ones who don’t look as sick? Frequently elderly, possibly with normal-ish vitals and no fever. Those can be a lot harder to spot, but they may indeed be septic. Also, for research purposes we have to have a common definition, so Sepsis 3 came up with something called the SOFA score.
The problem with the SOFA score is that its difficult to perform in the ED. It has parameters like bilirubin that often aren’t available when we want to screen out very sick patients. Fortunately there is the abridged version qSOFA, which identifies non-ICU patients who are at high risk of inpatient mortality.
So here it is, and if you get one thing from this episode, this is it:
There are ONLY 3 criteria to the qSOFA. 3 Criteria. RR > 22; AMS; SBP <= 100. That’s it. If you have two of these criteria, you are up to 14 times more likely to die of sepsis during a hospital admission. That’s pretty profound; these patients are very sick. This is meant to replace SIRS. It also captures a much sicker population than the patients included in the Medicare definition.
Jeff: So why do you think these parameters turn out to be so useful?
Jeremy: Drilling down into these criteria you can see the pathophysiology at work. Obviously, SBP < 100 means sick. Interestingly, an elevated RR also turns out to be prognostic, because you’re seeing the compensation for an underlying acidosis. WHen you see a patient breathing quickly, it’s either from a primary respiratory problem or them trying to relieve an underlying acidosis. The caveat here is that you have to check it. At our hospital in southern Manhattan, patients tend to breathe around 16. At our hospital in northern Manhattan they like to breath around 18. It’s probably because the air is thinner. Seriously though, you have to actually measure RR for this to work. Temp is not in QSOFA but we should be checking that too. And I mean checking it by putting something that measures temperature inside the patient. We’ve looked at the forehead and tympanic thermometers and in real world conditions, they tend to underestimate by a degree or more. Think about that. A patient with a headache and a temperature of 99.5 is a very different patient than one with a temp of 100.5. Make sure you measure temperature.
Nachi: Very true and these two patients can definitely go down very different management pathways! Rounding out our discussion on sepsis-3. We should note that severe sepsis is now a term of the past under Sepsis-3. And sepsis-3 redefined septic shock as “hypotension not responsive to fluid resuscitation” with the added requirement of vasopressors to maintain a MAP greater than or equal to 65 and with a lactate > 2. So quite a few changes!
Jeff: And Jeremy, sticky topic coming up here. Centers for Medicare and Medicaid Services (or CMS) quality measures - They haven’t really caught on to and adapted to Sepsis-3 yet, have they?
Jeremy: The CMS mandate is based on the presence of SIRS criteria. Sepsis 3 is based on SOFA. This is definitely confusing. Part of the challenge in discussing this topic is separating out the QI guidelines from what is actually relevant to patient care based on the latest evidence-based medicine.
Nachi: That seems fair. We’re really going to put you in an uncomfortable spot for a second and push you here Jeremy. Do you have any insight into why CMS isn’t interested in following the mountains of research that have led to sepsis-3? Is there a reason they are sticking to their current criteria?
Jeremy: I think some of it is the slow pace of bureaucracy and the time that it takes to develop a consensus on management. Even if we can agree on who is septic, it’s really hard, if not impossible to link the care to a pay-for-performance metric which is what CMS ultimately would like to see. That’s not how Sepsis-3, or for that matter, SIRS, was designed to be used. You’re trying to take a tool which was originally designed for research and mold them into a tool used for pay for performance.
Nachi: What a struggle. The CMS metrics are slightly different from the 2001 sepsis guidelines also. Take a look at Table 2 of the article for a quick comparison of sepsis-3, 2001 sepsis, and cms side-by-side. And for those on twitter, we’ll be sure to tweet this table out too for your review.
Jeff: With so many different scores and definitions, I think that adequately sets the stage for the challenge this month’s authors faced coming up with real evidenced based guidelines.
Nachi: Oh absolutely. And to make matters worse - this is a HUGE problem. We’re talking up to 850,000 ED visits annually in the US, and 19 million cases worldwide. Compounding this, sepsis results in death in approximately 1 out of 4 cases. Not only is it lethal, it is also very costly -- 17 billion dollars per year in the US alone!
Jeff: And don’t forget importantly the 30-day hospital readmission rate. Sepsis is coming in at a higher readmission rate and cost per admission than acute MI, CHF, COPD, and PNA.
Nachi: Let’s speak briefly on the etiology and pathophysiology of sepsis: we all know that sepsis is due to local infections that then become systemic. Previously, it was believed that the bacterial infection itself was the cause of the clinical syndrome of sepsis. However, we now know now that the syndrome of sepsis is due to the inflammatory and immunosuppressive mediators that were triggered by the infection. Normal immune regulatory safeguards fail and this leads to the syndrome. And interestingly, several studies have shown that critically ill septic patients experience reactivations of specific viruses that were previously limited to patients with severe immunosuppression.
Jeff: Definitely something to look out for in your critically ill septic patients. We should talk briefly about the most common inciting infections that lead to sepsis. In order, these are: pneumonia, intra-abdominal infections, and urinary tract infections. No surprises there!
Nachi: Yeah, that basically parallels my own experience, so that’s reassuring! That takes us to our next potentially controversial topic - blood cultures. Jeremy - we’re going to punt this one back to you
Jeremy: This is another interesting topic that has received plenty of attention. CMS loves blood cultures. It’s an easy metric to track. That doesn’t mean they’re always helpful. We looked at our patients with lactates between 2.1 and 4.0 which had “severe sepsis.” These patients were normotensive though, In other words, the ones that aren’t that sick. We found that blood cultures are useful about 20% of the time. That’s not bad. So what do we do? We draw cultures before pushing antibiotics. Is that helpful? Sometimes yes, does it waste money? Debatable. Does it help us meet our metrics, yes.
Jeff: And I think that gets at the crux of the problem here: we don’t want to delay antibiotics on anybody, but we must balance this with the potential harm of further increasing the drug resistant bacterial population via sound antibiotic stewardship. Remember also that there is a broad differential for sepsis, with several “sepsis mimics”. To name a few, we have PE, MI, CHF, acute pulmonary edema, DKA, thyroid storm, GI bleeds, drug intoxications, and withdrawal syndromes, just to name a few. In case that wasn’t enough check out Table 3 of the article.
Nachi: And we already mentioned the leading causes of sepsis, that’s pneumonia, intra abdominal infections, and uti’s. But remember the source can be anywhere. Be sure to also think of pyelonephritis, central line associated bloodstream infections, prosthetics, endocarditis, necrotizing fasciitis, and meningitis.
Jeff: I don’t think we need to dwell on this much longer - basically the differential is huge. Let’s move on to my favorite section - prehospital care.
Jeremy: 20 pages of evidenced based recommendations and your favorite is the prehospital section, what’s up with that?
Jeff: I’m an EMS fellow, what can I say… Anyway, on to my favorite section -- prehospital care. This is always a hot topic because the prehospital period is a special opportunity to get early interventions in for septic patients as 40 - 70% of all severe sepsis hospitalizations arrive via EMS.
Nachi: And in one study taking place in a large metropolitan area, prehospital care time was over 45 minutes, and less than 37% arrived with IV access. Of course, these numbers would vary significantly based on where you practice.
Jeff: So get this -- one study showed that out-of-hospital shock index and respiratory rate were highly predictive of ICU admission. So clearly early recognition and therapy may play a role here. Another study, however, showed knowledge gaps by advanced EMS providers in diagnosis and management of sepsis. And yet another study showed that only 18 to 21% of confirmed septic patients were suspected of having sepsis by EMS. Out of hospital fluids were started in only half of patients with severe sepsis. In essence, there is likely a strong role here for pre hospital protocols for identifying and treating sepsis.
Nachi: In terms of pre hospital treatments though, prehospital IV fluids haven’t been shown to improve mortality, but have been associated with shorter hospital stays. Prehospital sepsis protocols have been described, but in general more research is needed in this area.
Jeff: While prehospital care hasn’t yet been shown to improve the prognosis of septic patients, those presenting via EMS do have shorter delays to initiation of antibiotics, IV fluids, and early care bundles. EMS should focus primarily on stabilizing vital signs and providing efficient transport. If it’s possible to establish an IV and initiate fluids without delaying transport, EMS should do that as well.
Nachi: And of course, oxygen for the hypoxic patients! Moving on to history and physical for your presumed septic patient. Jeremy, what are the big hitting things here that you always ask and check for, and that you make sure your residents are doing?
Jeremy: After ABC’s and glucose, AMS is really important, it’s in the QSOFA SCORE. Unfortunately, this can be hard in many septic patients where they’re baseline mental status is less than perfect. The other thing is to try and find the source. Finding the source lets you make wise choices about therapy.
Jeff: Great point about the mental status - so many of our older population have an altered baseline, but recognizing changes from that baseline is key.
Nachi: Absolutely, with that in mind, let’s talk diagnostic studies, especially lactate. Where I trained, basically everybody was getting a lactate, even tired looking residents seemed to be having their lactates checked, and trust me, they weren’t looking that good...
Jeremy: Brace yourself: lactate is really important in septic patients. That being said, not every cause of elevated lactate is sepsis. There is this animal called Type B lactic acidosis can come from numerous drugs like albuterol. Just because you see elevated lactate doesn’t mean you can forget about the other causes. That being said, we know that patients with sepsis do better when they clear lactate.
Jeff: Seems like the evidence is definitely in favor of serial lactate testing…
Jeremy: For sure. At least until you have a reasonable trend towards improvement. We know lactate clearers do better. We’ve looked at our own lactate numbers. Interestingly, the takeoff point for sepsis seems to be around 2.5. Meaning that patients with altered vitals and lactates above 2.5 tend to do worse. But, there is a broad ddx to elevated lactate. What is true, though, is that lactate is a marker for badness. If your patient’s lactate is rising, yours should be too.
Nachi: I bet I’m a “lactate clearer”. I may add “lactate clearer to my CV,” sounds impressive. But I digress… Next up we have Procalcitonin. Since procalcitonin becomes elevated in those with bacterial infections, intuitively, this should be a valuable marker to assess in potentially septic patients. Unfortunately procalcitonin lacks negative predictive value so most literature supports its use in diagnosing pulmonary infections and for antibiotic de-escalation.
Jeff: Good to know, I’ve seen it being used a lot more recently and wondered how evidence based this test was.
Jeremy: Honestly, I don’t see Procalcitonin changing ED management at the moment. If you’re waiting for procalcitonin to start antibiotics or fluids, you’re waiting too long.
Nachi: Moving on, let’s talk imaging. Based on current studies, the authors recommend focused imaging only. In addition, they also note that our good friend, the point of care ultrasound, likely plays a role, as in one study, POCUS demonstrated a 25% improvement in sensitivity from clinical impression alone.
Jeremy: I think there are two ways POCUS comes in. One, lung ultrasound can be really useful to find that occult pneumonia or differentiating pneumonia from CHF. Two, your ultrasound is your best tool for assessing volume status. I try to look at the IVC of all my septic patients and echo them when possible.
Nachi: Right. So now we’ve examined, drawn labs and cultures, checked a lactate, may be obtained imaging… next up we should probably start treating the patient. Whether you like it or not, we have to discuss CMS.
Jeremy: Just to clarify before we start. CMS defines “severe sepsis” as SIRS + infection with a lactate of 2.1-4.0. Septic shock is SIRS + infection with hypotension or a lactate > 4.0. That’s where we’re at.
Jeff: Good point. Back to treatment: within the first 3 hours, for any patient with sepsis and septic shock, you must measure a lactate, obtain 2 sets of blood cultures, administer antibiotics, and give an isotonic fluid challenge with 30 cc/kg to patients with hypotension or a lactate greater than 4. Then, within the first 6 hours, you must apply vasopressors to achieve a MAP of at least 65, re-assess volume status and perfusion, and remeasure a lactate.
Nachi: This begs the question - are these recommendations evidenced based? Jeremy….
Jeremy: I’m so glad you asked that . Let’s start with fluids. Patient’s need adequate fluid resuscitation. Interestingly there are 3 large RCT’s, PROMISE, PROCESS and ARISE, that compared a Rivers type bundle to usual care. Surprisingly, they showed no difference. But when your drill down into these 3 trials, you see that “the usual care,” now generally includes at least 2 liters of fluid.
Jeff: Ok, so it seems that there is some pretty good data to support a rapid fluid challenge of at least 30 cc/kg. But how do we determine who needs more fluids and how much more they need. There must be an endpoint to all of this?
Jeremy: Another million dollar question. 30cc/kg is probably a good place to start. How much is too much? I think we need to be smart about our fluids. Some patients will need less and some will need much more. So, I remind my resident’s to be smart about fluids. Sono an IVC, trend a lactate, follow a urine output, do a passive leg raise, even check JVP. I mean just because you haven’t seen a unicorn doesns’t mean they’re not real. Do something to monitor volume status.
Nachi: Very important. Put your ultrasound skills to work here. They’ll only improve as you practice more. Jeff, let’s get started on the ever important topic of antibiotics.
Jeff: Sounds good. Current guidelines recommend that broad spectrum antibiotics be administered within the first hour of presentation for those with sepsis or septic shock, ideally with blood cultures being drawn beforehand. In one study, every hour of delayed abx administration was associated with an 8% increase in mortality. Since this 2006 study, other studies have had mixed results - with studies showing increased odds of death with delays in abx administration and others showing only a benefit in those with septic shock with or without hypotension with no benefit to those without shock.
Nachi: In terms of antibiotic coverage - you need to consider the site of infection, local resistance patterns, the presence of immunosuppression, and the patient’s age and comorbidities. Table 5 of the article is very thorough and should be kept as a quick reference. Jeremy do you have any specific recommendations for our listeners on how we should approach antibiotic usage in the septic patient?
Jeremy: I like to think about antibiotics a little more simply than referencing a table. I ask a couple questions. Does my patient need MRSA coverage ? Does my patient need Pseudomonal coverage? If the answer is no and no, then narrow your coverage. You don’t necessarily have to use a bunch of Vanco, or a big gun antipseudomonal like Pip/tazo. Also, have a look at your local antibiogram. I can’t tell you how many times this changes prescribing habits for even things like simple UTIs. I’m going to stray into some controversial territory here. The benefits of sepsis protocols are measured one patient at a time, but the harms are only measured in the aggregate. What does that mean? CMS metrics have caused us to use to use more broad spectrum antibiotics. As a result, we’re seeing more resistance. My resident’s tell me to make it easy, give em VZ (that’s vanco/zosyn) and it kills me. Every time you put a Z-pack into the world a pneumococcus gets it’s wings. So think more about your antibiotics, and know your local biograms.
Jeff: That’s a great way to think about it, I fear I’ve given a lot of pneumococci wings during my training… Next we’re on to vasopressors. The data is pretty clear on this one - norepinephrine is the recommended first line vasopressor for septic shock. In numerous trials comparing Norepi to dopamine, NE was far superior, with dopamine increasing arrhythmias in one trial and associated with an increased risk of death as compared to NE in another trial.
Jeremy: So here’s a question I get all the time: How can I give Norepi without a central line. Let’s use Dopamine, its safe peripherally. Ok, so follow that through. We’re going to give a drug to increase blood pressure by constricting blood vessels, but don’t worry, it’s safe peripherally. What does that mean? It means it doesn’t work!! It doesn’t give much blood pressaure. Dopamine is a lousy pressor. It causes a lot of tachycardia, which is not what you want in failing septic hearts. So what do we do if we don’t have a central line? We start norepi peripherally into a large bore IV for the time it takes us to get a central line. That’s where the evidence is. There’s a mortality benefit to NE over dopaine in septic shock.
Jeff: Right, this month’s authors note peripheral pressors may be safe for brief periods in settings with close monitoring. While this is commonplace in some hospitals, others haven’t yet jumped on that bandwagon. I think it’s important to mention that this is becoming more and more commonplace, even in the prehospital realm. With the service I fly for, we routinely start peripheral vasopressors without hesitation. But this isn’t limited to the air. Many ground 911 services have also adopted peripheral vasopressors in a variety of settings.
Nachi: I’m sure there are many trials to come in the future documenting their safety profile, but moving on to the next pressor to discuss... vasopressin. This should be your second line vasopressor for septic shock. In the VASST trial, low-dose vasopressin was found to be noninferior to NE. In other trials, vasopressin also appeared to show a potential benefit in those with AKI and sepsis, although the subsequent VANISH trial (perhaps the best name for a clinical trial so far) failed to demonstrate a benefit to vasopressin titration with regard to renal outcomes in septic shock. Vasopressin has also been shown to reduce NE dosing when administered at a fixed dose of 0.03-0.04 units/min.
Jeff: Next we have epinephrine. In one study epinephrine and NE were equivalent in achieving MAP goals in ICU patients with shock, however several of those receiving epi developed marked tachycardia, lactic acidosis, or an increased insulin requirement. The increasing lactic acidosis could confound the trending of lactates, so in those requiring inotropy in addition to some peripheral squeeze - the authors recommend adding dobutamine to norepinephrine instead of starting epinephrine. Although, keep in mind, this can lead to some hypotension so remember to start at low doses.
Nachi: Phenylephrine, a pure alpha adrenergic agent, is next and should be considered neither first nor second line, but it may have a role as a push dose agent while preparing other vasoactive agents.
Jeff: And lastly, we have angiotensin 2. One recent 2017 study examining the role of angiotensin 2 in those with septic shock already on 0.2 mcg/kg/min of NE found that those receiving AT2 had significant improvements in MAPs as well as cardiovascular SOFA score at 48h with no difference in mortality. Unfortunately, these benefits do not come without risk as AT2 may increase risk of arterial and venous thrombosis and potentially thromboembolism. Clearly, one study isn’t enough to change practice, but it’s certainly food for thought.
Nachi: So that wraps up vasopressors. Jeremy, we’re on to corticosteroids -- another hotly debated topic. When do you give steroids in sepsis?
Jeremy: Hmmm steroids, this is an age old question. No study has clearly supported the blanket use of steroids in septic shock. Several like CORTICUS and ADRENAL showed no difference. I will use hydrocortisone for pressor refractory shock. Meaning, you’ve tried everything else, so you might as well try. Also, I do tend to avoid Etomidate, given the possibility of adrenal suppression and that there are several other induction agents, notably Ketamine that don’t have this problem.
Jeff: Those trials are certainly important, thanks for bringing them up - Especially with all the FOAM content out there, it’s incredibly important to look back at the data to understand where certain recommendations are coming from. Anyway… one quick note on blood transfusions before we move on to special populations - Although part of the original early goal directed therapy, thanks to data from the TRISS trial which showed no difference in outcomes with a transfusion goal of 7 vs 9, transfusions are reserved for those with a hbg of less than 7.
Jeremy: One population we should make sure to mention and be careful with is end stage liver disease. In the ER, we tend to miss SBP alot. Mostly because these patients have lots of reasons to be sick and they already have elevated lactate because of their deceased clearance. My practice is to give a dose of Ceftriaxone and sent a diagnostic tap to patients who are sick and have ascites.
Nachi: Alright Jeremy, let’s talk controversies in sepsis. We’re giving you all the big questions this month!
Jeremy: We’ve already talked about fluids and how much to give. Just a reminder that a history of CHF doesn’t preclude proper fluid resuscitation. I think broad spectrum antibiotics for relatively well patients is a big controversy. Our national rates of antibiotic resistance are terrible, and yet we’re using more antibiotics all the time. There are very few if any antibiotics coming down the pharma pipeline and we’re going to have to face the music eventually. Finally, we need national metrics that mirror clinical evidnece. Protocols should be a tool and not a crutch. You know what’s best for the patient in front of you, so don’t let metrics or protocols make you do things you think are not in your patient’s best interest.
Nachi: So how do you escape the hospital protocols and CMS and do what’s best for your patient without “getting in trouble”?
Jeremy: Here’s how I deal with it as the one who reads and QI’s all of our sepsis charts. I tell my colleagues to do what’s right, and if you need to deviate from the protocol tell me why. As long as you can explain your decision, I’ll support it. Explaining your thinking is good clinical practice and is good medico-legal practice. CMS has been unable to link these metric to payment, simply because no hospital can meet them with any regularity. It’s important that we advocate for our patients or nothing will change. Make them respect you for the highly educated professional that you are, and your patients will ultimately benefit.
Jeff: Preach!! And before we close out with disposition, there are a few new therapies and trials on the horizon to keep a lookout for. The RACE trail examined the role of L-carinitine. The VICTAS trial is looking at vitamin C, thiamine, and steroids in sepsis. The CLOVERS trial is looking at early vasopressors vs a crystalloid liberal strategy. And lastly, IL-7 is also being investigated. All really cool stuff that could change how we manage sepsis in the future..
Nachi A few quick notes on disposition before we close this episode out. Certainly not all patients meeting SIRS require admission, but many do. Those with qSOFA of 2 or higher represent a sick population and an ICU admission should be considered. Even for those with a qSOFA of 1 but a lacate over 2 -- they have a mortality approaching that of patients with a qSOFA of 2. Be careful just sending a patient who is on the fence to the floor because several studies have demonstrated that patients who are later upgraded have worse outcomes.
Jeff: That’s in line with the general themes we’ve laid out today - definitely better to start early with aggressive care rather than play catch up later. Jeremy - in 30 seconds or less, what are the most salient points in the management of sepsis that you would like our listeners to take with them from this episode.
Jeremy: Here are my take aways:
qSOFA, RR, AMS SBP < 100
Norepi, not Dopamine - it doesn’t work!
Be smart about fluids!!
Be smarter about antibiotic use!
You are the best advocate for your patient, despite what anyone else says!
Jeff: Excellent, so that wraps up the October 2018 episode of Emplify. A big thanks to Jeremy Rose for joining us.
Jeremy: Thank you for having me!!! It was great talking with you.
Nachi: For our listeners -- additional materials are available on our website for Emergency Medicine Practice subscribers. If you’re not a subscriber, consider joining today. You can find out more at www.ebmedicine.net/subscribe. Subscribers get in-depth articles on hundreds of emergency medicine topics, concise summaries of the articles, calculators and risk scores, and CME credits. You’ll also get enhanced access to the podcast, including the images and tables mentioned. You can find everything you need to know at ebmedicine.net/subscribe.
Jeff: And the address for this month’s credit is ebmedicine.net/E1018, so head over there to get your CME credit. As always, the ding sound you heard throughout the episode corresponds to the answers to the CME questions.
Nachi: Lastly, be sure to find us on iTunes and rate us or leave comments there. You can also email us directly at firstname.lastname@example.org with any comments or suggestions. Talk to you next month!