Early this year, a publication in The Lancet analysed the global, regional and national incidence of sepsis-related mortality (Rudd et al, 2020), which is an important step forward in sepsis research. Prior to this study, our sense of the scale of sepsis incidence has always been extrapolated from hospital data in wealthier countries, and has usually excluded children or only covered a limited time period, making the data difficult to generalise. The lengths to which our esteemed colleagues stretched to compile, analyse and interpret the data they gathered is an impressive undertaking for which each and every clinician globally owes them a debt of gratitude.
In the UK, however, the findings of this study are in danger of misuse in popular media, and potentially even in the clinical context as health professionals discuss sepsis with patients and their families. This brief comment will attempt to address this concern, so that the relevance of Rudd et al (2020) is demystified for clinicians and patients.
About the report
Utilising data from the Global Burden of Disease database, the report analyses all causes of mortality from the period between 1990 and 2017—I am sure that readers will immediately appreciate the huge data set involved. The data are drawn from multiple sources, rather than only in-hospital clinical records, which improves its generalisability and validity. The data have been modeled extensively to define cases of implicit and explicit sepsis-related death, and then segmented by other epidemiological factors for the purposes of analysis and interpretation. This simple summary does not do such a monolithic effort justice, but the report gives us for the first time a global overview of the extent of sepsis-related mortality in multiple populations and in relationship to other human characteristics. The chief goal of the report, which seems to have been well achieved, is to estimate ‘the global, regional, and national incidence of sepsis and mortality’ (Rudd et al, 2020). Their chief findings were that, for the time period analysed:
Is sepsis a cause of death?
The authors of the article are at pains to properly nuance the relationship between sepsis and death from infectious causes, in a manner that clinicians and the media run the risk of failing to honour. The authors state categorically in fact that:
‘Since sepsis is presumed to result from underlying infection, it is inherently an intermediate cause of health loss… intermediate conditions reported as the cause of death are considered miscoded.’
This study analyses the incidence and mortality of sepsis-related deaths, and the authors themselves use this vernacular throughout. As far as our understanding at the moment goes, patients do not ‘catch sepsis’, or spontaneously ‘become septic’ without another underlying cause, namely infections, some of which may themselves result from distal causes such as injury. In light of this, headlines such as ‘one in five deaths due to sepsis’ (BBC, 2020) are supremely unhelpful, and misrepresent the findings of this study. Members of the public are given the impression by this kind of scaremongering that medical systems globally are failing to diagnose or treat sepsis, as if it were a mysterious, discrete disease, which it is not—it is a dysregulated response to an infection which compromises organ function.
An example to try to clarify: imagine if the BBC suddenly pronounced, ‘Thousands dying of cardiogenic shock’. Members of the public hear a headline like this and presume that cardiogenic shock is some sort of newly discovered disease that hospitals are currently unable to diagnose, resulting in people very suddenly dropping dead of this mysterious illness. Clinicians understand that cardiogenic shock is a clinical state through which an unstable patient with cardiac illness may pass prior to death or recovery, but it is not a standalone problem: it is secondary to a cardiac problem. In the same way, perhaps it would be better for us to think of sepsis as a clinical state through which an unstable patient with infectious illness may pass prior to death or recovery. Perhaps sepsis is simply an inevitable part of the dying process from infectious aetiology—how infection can kill, rather than what kills us.
Can paramedics—or anyone—diagnose sepsis?
Sepsis, if thought about as a physiological state, is something about which a clinical judgement is made based upon history-taking and examination findings. In the prehospital setting, we typically look for signs of underlying organ dysfunction, with the presence of a suspected infection. In-hospital diagnosis is not dissimilar, but usually also includes blood test evidence of both infection and organ dysfunction to shore up clinical suspicion (e.g. arterial blood gases, lactate, full blood count, other inflammatory markers).
Even with these laboratory tests, however, diagnosis of sepsis is still considered a ‘clinical’ diagnosis—that is, a diagnosis we make based upon a combination of findings because there is no definitive investigation. There is no ‘sepsis test’. So in one way, yes we can diagnose sepsis, even prehospitally, and initiate treatment; but in another sense, no, we cannot diagnose sepsis, not definitively—and nobody can.
The Lancet article addresses this ambiguity in infection-related deaths by accounting for ‘implied sepsis’ in mortality where sepsis is not explicitly mentioned in case reports, and the fact that they do so is interesting because it means we're now making the assumption that death from infection almost always involves sepsis, which it may not. A patient may contract an infection which causes complications to arise, resulting in death prior to a septic state occurring; for example, arrhythmia and cardiac arrest following myocarditis (Lee et al, 2006).
In paramedic practice, our clinical suspicion of sepsis should remain high when dealing with patients who feel themselves unwell enough following an infection to call 999. You can diagnose and treat sepsis as a paramedic, and you are expected to do so, by both the Health and Care Professions Council (HCPC) (2014) and other bodies such as the College of Paramedics (2020).
One in five deaths globally?
The article nuances its conclusions significantly in a way that news headlines do not. Firstly, sepsis-related deaths from the year in which this figure was drawn (2017) include sepsis following several non-infectious causes—that is, sepsis following the presence of infection and organ dysfunction that has developed from another primary cause of illness, such as road traffic accidents, neonatal disorders and maternal complications—which is something media articles fail to mention. In addition to this, overall incidence of sepsis-related mortality and morbidity has fallen year on year according to this study. According to the modelling of the data presented, sepsis is happening more than we thought it was last year, but less than it actually was the year before.
Sepsis is involved in around one in five deaths globally, but it is not causing one in five deaths.
Cause of death is a multifactoral issue and single-item headlines are unhelpful in understanding global mortality. Arguably, if we want a single-item headline that covers the largest cause of death globally, it would have to be economic disadvantage. Somehow, I suspect we will not be seeing a BBC news article any time soon that reports, ‘4 in 5 deaths due to poverty’.
How should this study influence my practice?
Several findings from this study can help improve practice for UK paramedics. Most directly, we should remember that sepsis is both high incidence and high mortality. Some of our patients will already be septic on our arrival. Some will become septic at a later point. Sepsis presents a real and substantial risk of death for these patients, and it is an emergency. Following local and national guidelines for early administration of fluids even in normotensive patients is important, as fluids may have dilutional benefit for circulating inflammatory mediators and help improve oxygen delivery to tissues, rather than just to improve blood pressure readings in hypotensive patients against arbitrary systolic cut-offs.
Secondly, we should be aware that sepsis is more common in patients at extremes of age (where immunity is naturally worse), in patients with diarrhoeal disease, and in patients with lower respiratory infections. These are important risk factors to bear in mind when we are considering how quickly we need to convey a patient to hospital, whether to pre-alert the receiving emergency department, and whether or not we should be treating for sepsis alongside treating for other primary diagnoses (such as suspected meningococcal disease) during transport. Alzheimer's disease, urinary disease, chronic obstructive pulmonary disease (COPD) and diabetes are also common risk factors for both incidence and mortality.
Thirdly, we need to be in touch with signs and symptoms associated with organ dysfunction, rather than just signs and symptoms of infection. Somebody presenting with pyrexia but otherwise well appearing is rightly considered low risk. However, altered mental status, haemodynamic instability, an altered rate or pattern of respiration are signs of an altered physiological state, which when associated with recent infection should be considered ‘red flags’ for sepsis—as indeed we already consider them to be in most UK ambulance services. Continued professional development in physical examination may help us to make discriminatory decisions about risk for those less obvious cases, however, where we may be the first clinicians to assess the patient and decide the level of risk involved in their condition.
Fourthly, these data actually show us, crucially, that cases of sepsis are falling and more people are surviving sepsis than ever before. This outcome of this report, read rightly, is that although sepsis is still involved in a great deal of global mortality, things are improving. Therapeutic interventions, improvements in practice and early recognition of sepsis risk all have a role to play in this, and in day-to-day paramedic practice on frontline ambulances.
Fifthly and finally, in our communication with patients, we should remain transparent about our concerns but avoid unnecessary scaremongering. The word ‘sepsis’ has become laden with media-fuelled misconception, and we might be better talking to our patients about ‘a progressing infection’, ‘an infection your mum/dad/partner is struggling to get on top of’, or other more understandable representations of suspected sepsis. Kindness to our patients and families costs nothing while we are treating them and their loved ones.