Is ‘sepsis’ a failed paradigm?

Abstract

In this issue, Sabir et al report a detailed description of a representative series of patients identified as ‘septic’ in their emergency department. Many expert bodies have called for a better understanding of the epidemiology of sepsis and the authors are to be commended for this work. They included patients who had blood cultures taken for presumed sepsis. Their main finding is the substantial burden of chronic disease and comorbidity that many of these patients were living with before they required emergency department treatment. It is striking that less than half were living at home, and nearly 90% had some important comorbidity. Furthermore, only 6.5% were referred to intensive care. Few of these patients presented in septic shock. These patients were sick, with a 13.2% inpatient mortality, but it is difficult to escape the conclusion that many were simply at the end of their life. It is possible that ‘sepsis’ has taken over from pneumonia as the ‘old man’s friend’, a term that medicalises death. The authors acknowledge the limitations of this work; the retrospective and singlecentre design and their case definition, but the findings have high face validity. For several years, there have been vigorous and wellmeaning campaigns to simplify recognition of abnormal physiology, and then institute aggressive treatments. In 2019, The UK Secretary of State for Health tweeted that ‘sepsis kills over 52 000 every year—each death a preventable tragedy’—a comment that has been criticised for fueling ‘sepsis hysteria’. Sabir’s data question this hyperbolic and simplistic approach and add useful evidence to recent highprofile discussions about sepsis hysteria. Furthermore, a recent metaanalysis found little benefit of antibiotics within 1 hour, compared with one to 1–3 hours. Many patients in Sabir’s series are likely to not be recruited to or excluded from randomised controlled trials, and it is uncertain whether many in this patient group would benefit from aggressive treatments. For instance, in the metaanalysis of the three landmark sepsis studies (PROMISE, ARISE and PROCESS), the median age of participants was 66, in Sabir’s paper, the patients are 10 years older. Internationally, many respected societies have recently called for less urgency and more evaluation of early antibiotic administration. The American College of Emergency Physicians and the Infectious Diseases Society of America have both recently advocated very short time scales for antibiotic administration only, where the patient is in septic shock. 7 In addition, it is well described that sepsis, among other conditions, is harder to diagnose in the elderly. Furthermore, many patients, even before the viral pandemic, with abnormal NEWS scores and ‘sepsis triggers’ turn out not to have bacterial infection, with over 30% of antibiotic prescriptions deemed inappropriate. 9 Identifying sepsis is difficult, both prehospital and in the emergency department. 10 There are undoubtedly some rare and serious conditions that present with sepsis type symptoms, such as meningitis and necrotising fasciitis and identifying and treating these early will improve outcomes. Likewise, treating sepsis in patients with surgical pathology improves outcomes. However, these modifiable conditions are in the minority. In this dataset, the majority of infections occurred in the chest, urinary system and skin. Is it really justified to put all of these conditions together under a banner of sepsis? Sabir’s and other data suggest that simply using sepsis triggers to initiate treatment with antibiotics leads to large numbers of patients being over treated for uncertain benefit. This has an opportunity cost for other patients, exposes these patients to iatrogenic risk and compromises antibiotic stewardship. Sepsis is a multisystem consequence of an infectious process that usually starts in a single organ, should we insist on making a diagnosis of the location of infection alongside sepsis? For instance, chest infection with sepsis, appendicitis with sepsis and so on. Such an approach would require more thought, but would describe the patient better and allow a more targeted approach. Should we reserve early antibiotics, without the benefit of laboratory results, only for cases of septic shock? In summary, Sabir’s report describes the realworld experience of managing sepsis in a UK emergency department that differs from the view presented in randomised trials and advocacy campaigns. In doing so, it raises some important questions for all of us around considering the balance between blunt protocols based on messy and conflicting evidence and doing what is right for the patient in front of us and for the wider population.