Defining Sepsis Then & Now, Part 1

ILLUSTRATION: Chris Whissen PHOTOS: shutterstock.com

Epidemiology 1979–2000

Before 2000, solid epidemiology data on sepsis were lacking. No formalized, generally accepted definition existed, even for septicemia, which was more commonly used. Additionally, prior to 2002, diagnostic codes for sepsis, severe sepsis, and septic shock were nonexistent.

Martin et al estimated that there were more than 10 million septic patients during a 22-year span, increasing 9 percent annually from 164,000 cases to 660,000.² Angus et al estimated 750,000 cases, representing more than cases of breast cancer, colon cancer, and AIDS combined. With a reported 500 deaths per day, mortality for sepsis paralleled that of out-of-hospital myocardial infarction, costing $16.7 billion nationally.³ Mortality estimates for the period ranged from 18 to 75 percent, depending on illness severity and population.^2–5 A meta-analysis reported mortality as 47 percent from 1991 to 1995 and 29 percent from 2006 to 2009, representing a 3 percent annual reduction across studies.

Most sepsis occurred external to the ICU. Approximately 60 percent of the Angus et al cases were identified outside of the ICU. In a global study, 88 percent of septic patients were identified outside of the ICU (56 percent in the emergency department, 32 percent on the ward, and 12 percent in the ICU).⁶

2001–2004

Wang et al estimated that there were 571,000 suspected community-acquired severe sepsis cases presenting annually to emergency departments. Mean ED length of stay was almost five hours, with more than 20 percent of the visits lasting in excess of six hours.⁷

2009–2011

Revised estimates total 850,000 ED sepsis visits per year. This constitutes one out of every 120 ED patients. More than 70 percent of septic ED patients were admitted, with 34 percent requiring ICU admission. However, almost half had an ED length of stay of more than four hours, and more than 10 percent stayed more than eight hours.⁷

Gaieski et al demonstrated how sepsis prevalence and mortality rates change depending on the definition applied. However, what is clear is that the majority of sepsis cases are identified and initial resuscitation for them is completed outside of the ICU.⁸

“And yet, at the present time, the subject is by no means fully elucidated, and it is not even possible to give a general definition of the term septicemia, which could currently represent all the different conceptions of its nature … at the present time.” —W.W. Van Arsdale

Sepsis Definitions 1990s

Sepsis definitions have evolved over time (see Table 1). The 1991 American College of Chest Physicians/Society of Critical Care Medicine conference definitions were based on a definition Bone et al decided upon when the investigators realized they needed to identify severe sepsis patients early in their course, prior to traditional culture availability, for a trial of high-dose methylprednisolone.^10,11 They defined the pragmatic, inflammatory criteria with organ dysfunction as “sepsis syndrome.”¹² Despite apparent face validity, problems with applicability resulted in dissatisfaction and prompted the 1992 consensus conference.¹⁰ Conference participants considered infection to be associated with microbial penetration and sepsis to be associated with the clinical, inflammatory response. Within this framework, the terminology of systemic inflammatory response syndrome (SIRS), sepsis, severe sepsis, and septic shock was developed. However, SIRS was criticized for lack of specificity and imperfect sensitivity.¹² Additionally, the process was criticized for predominantly involving North American representation.^13,14

(click for larger image) Table 1: Modern Sepsis Definitions<br>AMS: altered mental status, SIRS: systemic inflammatory response syndrome, ACCP: American College of Chest Physicians, SCCM: Society of Critical Care Medicine, MODS: multiple organ dysfunction syndrome, ATS: American Thoracic Society, ESICM: European Society of Intensive Care Medicine, SIS: Surgical Infection Society, *Presumed or identified infection

(click for larger image) Table 1: Modern Sepsis Definitions
AMS: altered mental status, SIRS: systemic inflammatory response syndrome, ACCP: American College of Chest Physicians, SCCM: Society of Critical Care Medicine, MODS: multiple organ dysfunction syndrome, ATS: American Thoracic Society, ESICM: European Society of Intensive Care Medicine, SIS: Surgical Infection Society, *Presumed or identified infection

2001

The 2001 International Sepsis Definitions Conference involved five medical professional societies. A stratification system, similar to cancer staging, was developed predominantly to predict therapy response rather than predict mortality.¹⁵ Components of the PIRO model were predisposition (baseline factors influencing survival, etc.), insult (infecting pathogens, source control, etc.), response (physiologic response, SIRS, etc.), and organ dysfunction (number of failing organs). However, the model was complex and difficult to standardize and quantify, and it was never universally applied. Both of the early definitions were based primarily on consensus of expert opinion.

2016

The Sepsis-3 definition focused primarily on infection and end-organ dysfunction. Sepsis became infection and worsening or new end-organ dysfunction based upon the Sequential Organ Failure Assessment (SOFA) scale or quick SOFA. Sepsis includes vasopressor-dependent hypotension and excludes lactate. Septic shock was defined as vasopressor-dependent hypotension with lactate elevation. The Sepsis-3 process made several important advancements. It would be the first definition conference to: 1) employ data analytics in conjunction with clinical expertise rather than solely relying on expert opinion, 2) attempt a standardized process, and 3) acknowledge continued refinement as science and our understanding evolves.¹⁶ However, the Sepsis-3 definitions have not been universally adopted by emergency medicine, critical care, and low- and middle-income countries.^17–24 The Surviving Sepsis Campaign (SSC) ultimately decided upon an intermediate definition, where sepsis was defined by the previous severe sepsis definition (not including persistent hypotension) and septic shock remained any vasopressor-dependent hypotension. Henning et al did an excellent job illustrating implications of the different definitions.²⁵

Unfortunately, sepsis continues to be an elusive entity, recognized by most and unmistakably defined by none.

“There are two great shocks for every emergency medicine resident: one, not every patient is sick, and two, many patients are much sicker than they first appear.” —Peter Rosen¹

Definition Development

Definition usefulness is based upon the purpose for which it is derived, combined with the ability to effectively operationalize it within a target environment. Purpose is based upon the context of value and prioritization.¹⁶ Effective operationalization involves pragmatic implementation within the environment to which it is applied.

“Emergency medicine is in the business of sensitivity.” —Donald Yealy

Importance of Perspective

When working in the ICU, the intensivist is concerned about specificity. There are limited resources appropriately reserved for the most acutely ill. If intensivist assessment of a patient denied ICU admission is incorrect, the patient is invariably still in a hospital-based, monitored setting. With unexpected deterioration, accommodations can be made and resources leveraged. In the emergency department, the physician is concerned about sensitivity, or who could be missed. An incorrect assessment of a patient who decompensates after being discharged home is not afforded the same accommodations and may have significant consequences. For example, when the SSC data from the septic shock manuscript were divided into hypotension and normal lactate or normotension and lactate >4 mmol/L, the mortality was essentially the same at 30 percent. Hypotension was included in the definition of sepsis, and lactate was not. Although the authors never dissuaded using lactate because, from their perspective, lactate added no additional predictive validity for mortality or ICU resources, it was only retained during hypotension. From an emergency medicine perspective, a quick point-of-care test that identifies a patient at risk is of great value. If mortality and a prolonged ICU stay are averted, it could signal a job well-done through early identification and treatment rather than unnecessary information.

If definition usefulness is based on priorities, it seems important to know what is valued the most by the clinicians who are implementing the definition.

“Labeling already differentiated ICU patients is very different than sorting, or triaging, undifferentiated patients in the ED or on hospital wards.” —David Gaieski and Munish Goyal

Surviving Sepsis Campaign

ACEP’s involvement with the SSC began in 2003, when an emergency physician who had just completed a critical care fellowship was in Amsterdam taking the European critical care boards at the European Society of Intensive Care Medicine annual congress, as the US boards were not an option. While there, the physician went to the SSC presentation in an auditorium filled with hundreds of intensivists and listened to the Phase II and III plans for this international critical care collaborative. During the question period, the emergency physician asked a simple inquiry: “If 50 percent of sepsis in the United States presents through the emergency department, would it be of benefit to have emergency medicine involved?”

A proposal was drafted, and conversations with the ACEP Board of Directors began through Arthur Kellermann, MD, MPH, FACEP. A lecture with panel discussion was arranged for the ACEP Scientific Assembly. In a room accommodating 200 people, approximately 25 people attended that inaugural lecture. There were no other sepsis talks at ACEP that year.

Sepsis didactics became a standing component of Scientific Assembly. An article documenting ACEP’s alliance with the SSC and a letter from the ACEP president regarding ACEP’s involvement were sent to 35,000 members. During this time, sepsis mortality was high, with an estimated 500 deaths per day in the United States alone. As part of Phase I, Graham Ramsay, MD, conducted a survey of more than 1,000 physicians from six countries and found fewer than 17 percent of physicians agreed on a common definition or a standard treatment. In 2003, ACEP participated with representatives from 11 international medical professional organizations in the creation of the first set of sepsis management guidelines advising a standardized management approach to sepsis. Guided by Phillip Dellinger, MD, there was meaningful collaboration. During periods of intense academic discussion, where few agreed on how specific data were interpreted, he would carefully consider the different perspectives and then skillfully find common ground. ACEP contributed meaningfully with 100 percent of the major content revisions and 75 percent of moderate content revisions accepted. Through Phase III, headed by Mitchell Levy, MD, from the SSC, and Sean Townsend, MD, from the Institute for Healthcare Improvement (IHI), an innovative approach of structured, data-driven implementation was taken. Although there were periods of academic differences of opinion on specific elements of the guidelines, the representatives generally came to a consensus and worked collaboratively with mutual respect.

Over the past two years, simultaneous with the departure of Dr. Dellinger from the SSC, the collaborative culture changed. Although the majority of sepsis patients are identified and initially treated outside of the ICU, clinical perspectives from service lines such as emergency medicine, hospitalists, nurses, or infectious disease were not considered when developing definitions. This decision resulted in reduced uptake. Additionally, a few months after completing a yearlong process of guideline development, SSC executive committee members, independent of the multispecialty guideline committee, published the Hour-1 Bundle. This bundle, directed toward ED practice without representation from the emergency medicine community and without review by the multispecialty SSC Guidelines Committee, is misguided and should serve as an alarming precedent for all professional organizations participating in the SSC. When a letter to the editor was written in response to the Hour-1 Bundle, it was not published due to a Society of Critical Care Medicine (SCCM) policy of having concerns directed to the guideline committee chair.

This is a challenging time. ACEP is working diligently with SCCM and the SSC. Although individual service lines can incrementally impact mortality, no service line can truly optimize sepsis survival alone. From 2003 to 2015, patient survival has improved due in part to this global cooperative work. I hope we can get back to the real mission, caring for the patients we are called upon to serve. Our patients need us all to be more than individual service lines; they need us to provide a service.

Part 2 of this history will appear in the December issue.

Complete the CME activity.

Dr. Osborn is professor of surgery and emergency medicine at Barnes-Jewish Hospital/Washington University in St. Louis, Missouri.