Secondary Logo

Emergentology: New Sepsis Bundle is Built on Good Intentions and Selection Bias

Walker, Graham, MD

doi: 10.1097/01.EEM.0000546142.79545.18

Dr. Walker is an emergency physician at Kaiser San Francisco. He is the developer and co-creator of MDCalc (, a medical calculator for clinical scores, equations, and risk stratifications, which also has an app (, and The NNT (, a number-needed-to-treat tool to communicate benefit and harm. Follow him on Twitter @grahamwalker, and read his past columns at



The new Sepsis-3 definition is ill-defined without easy, agreed-upon clinical criteria, and, as I noted last month, we now have one hour from ED registration to figure out who is septic, which can be very subtle, and to measure lactate, order blood cultures, start rapid IV fluids, give broad-spectrum antibiotics, and administer pressors for hypotension. (“Lies, Damned Lies, and Sepsis Bundles,” EMN 2018;40[8]:26;

Many in the emergency medicine and critical care communities began to push back, with Scott Weingart, MD, and Josh Farkas, MD, from EMCrit and PulmCrit even organizing a petition to retire the SSC guidelines. ( So I decided to talk to Mitchell Levy, MD, and Laura Evans, MD, the intensivists who authored the update.

Drs. Levy and Evans seem like intelligent, reasonable people who are trying to improve the care of sepsis patients. Unfortunately, sepsis has been poorly defined, and their selection bias as intensivists limited their ability to develop a bundle without significant unintended consequences for everyone else who presents to the emergency department.

They explained that the bundle was set to one hour because the quality measures for STEMI, stroke, and trauma—door to balloon, door to drug, the golden hour of trauma—made hospitals prioritize and optimize care. “It forces EDs to try to improve their processes up front,” said Dr. Levy.

The thinking is that setting a short time for sepsis interventions forces the lab to be more efficient, triage time to go down, and radiology to do the chest x-ray faster. If hospitals were to do this on their own, we wouldn't need this one-hour timeline, they said. Drs. Evans and Levy said mandating an unrealistic, time-dependent surrogate marker like 60 minutes will improve outcomes in sepsis patients by making everyone do things faster.

This is news to me that we EPs are not prioritizing sepsis patients and are somehow giving them short shrift. The bundle guideline even stated that “sepsis is a medical emergency,” and recommends that “clinicians begin treatment immediately, especially in patients with hypotension, rather than waiting or extending resuscitation over a longer period.” (Find me an emergency physician or nurse who allows a patient to remain hypotensive without evaluation or intervention for a prolonged period time.)

I unsuccessfully tried to pin down the authors on how we should define or identify patients with sepsis. Dr. Levy said “this bundle is meant to apply to patients with sepsis and septic shock, according to Sepsis-3 definitions.” I asked him how this should be done in patients who present without slam-dunk sepsis or septic shock. What about the flu patient with tachycardia, fever, and cough? Doesn't Sepsis-3 require us to do a SOFA score on them? An ABG, LFTs, a CBC, and a chem 7? They reversed course a bit, with Dr. Evans saying that “a lot of people have different ways of identifying patients with sepsis.”

Back to Top | Article Outline

No ‘Knife in Chest’

I pushed a bit more with the flu patient example, asking about giving broad-spectrum antibiotics to all flu patients within one hour of arrival and voicing concern about unintended consequences. Dr. Levy said one dose of antibiotics was not harmful and that if the treating doctors think there is “no evidence of bacterial infection, they can rapidly decrease the antibiotics.” Do they have any idea how many flu patients visit the ED each winter? Dr. Levy clarified, “If you truly think the patient does not have a bacterial infection, you should not have to follow this bundle. ... If the guy comes in saying he's coughing up yellow sputum and has a fever, however....” (Has Dr. Levy never met an ED patient with the flu who reports coughing up sputum?)

The intensivists really did seem like reasonable people. They acknowledged most of my concerns, admitting that many of my questions were realistic. None of their acknowledgements, however, made it into the bundle. There's no “out” written into the bundle allowing me to withhold antibiotics if I think a patient has a viral illness. They really wanted me to focus on the spirit of the bundle, not the letter of the bundle, by looking at the core messages. They explained that the bundle is a target, not a regulation. They said the current three-hour and six-hour sepsis bundles have a “compliance of 65 percent, and that will get you to the 95th percentile of hospitals.” It's weirdly clear to me that the Surviving Sepsis Committee understands that this sepsis stuff is hard—Drs. Levy and Evans seemed pleased with 65 percent compliance—but they're making recommendations that are even harder for EDs to achieve.

It's for this reason and several others that the Infectious Diseases Society of America refused to sign on to the Surviving Sepsis Campaign update in 2016. Reading their article of dissent is an EP's dream come true: A subspecialty group that agrees that “sepsis” and even “infected” are confusing and hard and that there are major consequences to trying to bundle care and give out antibiotics like candy! ISDA's concern was the guidelines' failure to recognize the practical difficulties clinicians face trying to diagnosis sepsis. Studies show that 40 percent of patients admitted to ICUs with a sepsis diagnosis do not have an infection. ( “We are fearful...that stipulating an aggressive, fixed time period ‘from the time of recognition’ may lead to unintended consequences,” the IDSA said.

I'll leave you with a statement from Dr. Weingart: “Metrics and time measurements work really well when you've got a single, obvious definition: ST elevations on an EKG, a patient not moving his left arm and leg, a guy with a knife sticking out of his chest. But there's no ‘knife in chest’ for defining sepsis. How are we supposed to follow this bundle in an hour if we don't even know who it applies to?”

If this bundle ever gets advanced to any quality level like the Centers for Medicare and Medicaid Services, the Joint Commission, or the National Quality Forum, we'd better be ready to speak loud and proud or we'll end up with an even more ludicrous practice mandate dictated to us by a few well-meaning intensivists who don't see all the flu, pneumonia, pyelonephritis, and cellulitis patients we want to send home appropriately and safely.

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.