The UK Sepsis Trust was founded by, and still receives senior leadership and strategic direction from, frontline clinicians passionate about improving outcomes from sepsis. Over the years, we have demonstrated leadership in this country and others in effecting systems change.
With this leadership comes responsibility. We are acutely aware of the intrinsic interrelationship between sepsis and antimicrobial resistance (AMR), and the potential societal impact which might result from getting the balance wrong. We’re committed to driving responsible improvement, to responding to new evidence and new ways of working, and to working collaboratively with agencies to ensure we deliver the care our patients need whilst minimising adverse consequences.
There is a lot we still don’t know about sepsis and severe infection, most particularly the lack of robust data on its incidence and outcomes. Whichever way you cut the existing data, however, sepsis is a huge problem. We have prepared below some responses to frequently asked questions to help you understand our mission, our values and our strategy for effecting change responsibly.
The first iteration of the International Surviving Sepsis Campaign (SSC) guidelines was released in 2004. At the outset, the Resuscitation Care Bundle, created from these guidelines by the Institute for Healthcare improvement in the US, incorporated a protocol known as Early Goal-Directed Therapy (EGDT). EGDT required central venous access, a liberal transfusion strategy and the use of vasoactive infusions. These interventions (and therefore the Bundle) were inaccessible to the majority of health professionals working outside Critical Care. The founders of the UK Sepsis Trust therefore set about simplifying these evidence-based guidelines by identifying those tasks that could be delivered by any junior health professional. The result of this was the Sepsis 6 – designed for use in Emergency Departments and on wards – which was launched in 2006 alongside an associated education program.
The Sepsis 6 is now used in over 30 countries across the world – and we believe much of its uptake is due to its simplicity and empowering nature.
We remain committed to responding responsibly to new evidence, as well as to changes in the way health care is delivered. In 2015, as we prepared for the launch of the NICE guideline on sepsis, we responded to increasing evidence in other disease states that hyperoxia may be harmful. We therefore amended the first element of the Sepsis 6 from a recommendation to use high flow oxygen to a recommendation toward targeted oxygen saturations. Similarly, in the years leading up to this, we amended recommendations around fluid challenges to be less liberal in light of increasing evidence of harm with over-resuscitation.
2019 has seen the most fundamental change to the Sepsis 6 since inception, as we are increasingly mindful of the interrelationship between improving outcomes from sepsis and antimicrobial stewardship. As a result, the first element now recommends that a senior health professional be consulted as soon as practicable in order that care can be de-escalated if appropriate and other causes of deterioration considered.
The second significant change is that the final element of the Sepsis 6, which used to recommend monitoring urine output, now also includes monitoring more generally, including the use of NEWS2 and serial lactates.
No. The principal behind care bundle methodology is that bundle compliance is considered as a whole rather than as individual elements – in other words, it’s important the whole process be completed. However, the Sepsis 6 permits – in fact encourages – variance where appropriate clinical rationale is applied. This may be best illustrated using examples.
If a patient is admitted following a three-day history of diarrhoea and vomiting with a mild febrile illness, and the Red Flag qualifying the patient for treatment is a low urine output, an experienced clinician might reasonably assume that this could simply be pre-renal renal failure as a consequence of dehydration from viral gastroenteritis. They may choose to appropriately withhold antibiotics whilst awaiting results of a differential white cell count, restoring circulating volume in the meantime.
Similarly a patient, who is known to have gallstones and is admitted with epigastric or upper abdominal pain and an acute abdomen, might have pancreatitis. It would be reasonable to withhold antibiotics, even in the presence of Red Flags, until the results of serum amylase and any imaging are available.
In both situations, providing the clinical rationale behind withholding antibiotics is documented, we would consider these aspects of the patients’ care exemplary. We would propose that any audit process should not record such clinical good sense as non-compliance with the Sepsis 6 bundle.
From 2001 until 2016, the international consensus definition for sepsis required the presence of two or more SIRS – the Systemic Inflammatory Response Syndrome – criteria as a consequence of infection. A diagnosis of severe sepsis – which we now consider to be sepsis – required the presence of one or more organ dysfunctions from a complex list of criteria.
SIRS was based upon a set of criteria originally proposed in 2001. SIRS is highly sensitive and very non-specific to either infection or sepsis. The majority of patients with SIRS as a consequence of infection, but without organ dysfunction, require neither hospital assessment nor intravenous antibiotics, and only a tiny minority require antibiotics with any degree of urgency.
The evidence around urgent intervention existed therefore only for those patients with organ dysfunction. Herein lay the problem: the list of dysfunction criteria was so complex that, within the UK, healthcare systems drifted away from its use. Care pathways within hospitals rationalised this decision-making to SIRS + Infection to be treated as sepsis, including the administration of broad-spectrum intravenous antimicrobials: this was clearly inappropriate.
In the absence of an existing pragmatic alternative, we therefore worked with NHS England’s patient safety team and representatives of the relevant Colleges to identify a set of criteria which were felt to identify a sick cohort of patients and act as a surrogate for a formal diagnosis of sepsis. Red Flag Sepsis was essentially borne out of those criteria from the then international definition which could be measured at the bedside, together with parameters from the NEWS score which carried an individual weighting of 3. All parties at the time agreed that this was a pragmatic way to identify a cohort which warranted urgent intervention.
From 2015 until 2016, our recommendation was that only those patients with 2 or more SIRS criteria as a consequence of infection AND one or more Red Flags be treated as sepsis. This was at odds with widespread more liberal practice at the time. The intent was of course to limit the use of antibiotics to those patients who needed them, rather than to drive increasing use. In time (see below), as more evidence-based alternatives come to light, we will embrace these and issue recommendations around them. From 2016 with the release of the Sepsis-3 definition, we changed from SIRS criteria as the screening prompt to NEWS2, continuing to recommend Red Flags as a treatment prompt.
Issued in 2016, the Sepsis3 definition redefines sepsis as a life-threatening condition in which the body’s response to infection causes injury to its tissues and organs. Organ dysfunction is a key component in any diagnosis of sepsis.
With increasing national adoption of NEWS2 as a standardised scoring system for acutely unwell adults in the UK, we agree with NHS England and other bodies that it should be the standard screening prompt for sepsis in adult patients.
Many patients with an aggregate NEWS2 score of 5 or above are critically unwell, and indeed national hospital mortality rates are in the region of 20%. It’s important to note, however, that this is an aggregate score. It is entirely possible to have a NEWS2 score of 5 without any evidence of organ dysfunction– a patient who is receiving 28% oxygen with a respiratory rate of 21 but with normal oxygen saturations, and a temperature of 38.3oC, is a good example of this.
We therefore recommend, in much the same way as we did not recommend treating patients with SIRS + infection as sepsis, that we don’t assume everyone with NEWS2 of 5 or above has sepsis.
To further qualify patients for urgent treatment therefore demands assessment beyond a NEWS2 of 5 or higher in the context of an infection. Where resources and skills allow, urgent assessment and competent decision-making by a senior clinician should trump any clinical tool and would act as the gold standard. Due to the time critical nature of sepsis in many patients, however, this is not always possible. Where an organisation cannot guarantee urgent assessment by a competent decision maker, we continue to recommend the use of Red Flag Sepsis as an empowering treatment prompt.
As soon as evidence-based, pragmatic alternatives to this strategy become available, we will embrace these and incorporate into our clinical pathways.
This was alleged in a letter to the Lancet by Mervyn Singer and colleagues in late 2019. Our response is currently under consideration for publication by that journal and will be posted in full here no later than 17:00 on January 16th 2020, whether or not it has been published.
There is general consensus that there is an intrinsic interrelationship between the three pillars of infection management: infection prevention, rapid treatment of time-critical infection and antimicrobial stewardship. It is essential that none be considered in isolation from the others. Just as it would be inappropriate to consider strategies to improve outcomes from sepsis without due consideration to AMR, so it would be inappropriate to fight AMR without considering its potential impact on patients with sepsis. We are facing an existential problem.
As such, all activities driving improvements in the way we recognise and manage sepsis need to be mindful of adverse consequence. Programs should evaluate the impact on total antimicrobial consumption, and the way in which the geography of antimicrobial prescribing is altered. Such considerations need to address the whole system rather than an individual location of care, as the interrelationship is likely to impact in unpredictable ways. An example can be seen in data from the Royal Pharmaceutical Society around antibiotic usage following the introduction of NHS England’s CQuIN commissioning incentive on sepsis. Whilst antibiotic consumption in Emergency Departments doubled in a few short years, total consumption across English hospitals remained largely static: it appeared that the result was that we were front-loading antibiotics without adverse consequence when considered across the system. We need to continue to monitor this situation closely if we are to get the balance right.
No, not all patients with Red Flag Sepsis (RFS) will have sepsis. Sepsis can be defined by the Sepsis-3 physiological criteria, or by its pathophysiological state. Within the group of patients who have RFS, a majority will have infection which may be time-sensitive. A smaller group will have true ‘sepsis’. Large data sets are required so that we can understand our patients with infection and infection mimics with more detail and better identify the patients who will benefit from early antibiotics and interventions such as the Sepsis 6.
The UK Sepsis Trust has a history of supporting such projects, and have actively supported a ‘big data’ pilot project in south-west England which hopes to inform and improve this decision-making.
In the future, we hope that patients will have individual risk profiles for sepsis based on physiology, point of care testing, their age, comorbidities, medications and other rapidly available information. This means we will be able to make patient-level decisions about the harms and benefits of antibiotic utilisation.
No, not all patients with Red Flag Sepsis (RFS) will benefit from urgent antibiotics. Within the group of patients with RFS, a small and currently unidentifiable cohort will benefit from antibiotics within one hour (as soon as practically possible). A further (much larger) group will have infection with a time-sensitive element where antibiotics delivered within 1 to 3 hours are important and probably effect morbidity, mortality and length of stay. Some patients identified with RFS will not have infections or will not have infection with a time-sensitive element. Unfortunately, the data are not yet available to identify such patients with the sensitivity and specificity that we would all like (see answer above regarding big data).
There is also a pragmatic operational component of the decision to recommend antibiotics within one hour: the timeframe of one hour allows the care provider time to consider the harm vs. benefits of antibiotics (and the Sepsis 6) but dissuades them from handing over patient care or postponing a decision to an unknown point in the future. Three hour care windows which may be suitable for some (unidentifiable) patients will often span handover times and other transitions of care and environments, and are impractical for implementation in an Emergency Department (especially in today’s stressed NHS). We believe the utilisation of RFS, with the latest version of the Sepsis 6 (with step one as senior involvement) represents the ‘best’ solution in what is unfortunately an information-poor decision environment.
It is critical that we come to better understand the cohorts of patients who will benefit from urgent antimicrobial therapy, in order that we can use these essential agents more responsibly. At present, the data are not available to support these decisions. We urgently need large scale patient-level datasets if we are to get the balance between early intervention in time-sensitive infection, and antimicrobial stewardship, right.
From April 2016 until the end of March 2019, English hospitals were incentivised to prioritise sepsis identification and management through a CQuIN (Commissioning for Quality and Innovation) incentive applied by NHS England. The standards applied, which included screening of patients with an elevated aggregate NEWS score for sepsis, urgent antibiotic therapy for those with Red Flags or other evidence of organ dysfunction and (during the final year) showing evidence of an antibiotic stewardship discussion within 72 hours, now form part of the standard contract.
As with any improvement initiative, including the Surviving Sepsis Campaign’s work, a benefit of this strategy was to shine a light on sepsis at the organisational level, ensuring that organisations began to evaluate the quality and effectiveness of their sepsis care at Board level and provided regular training to staff. It is undoubtedly true that the reliability of care improved – NHS England’s data showed that, among the more than three quarters of English hospitals submitting data, the reliability of screening patients with physiological derangement in the context of infection increased from a low baseline to over 85%, and the reliability of urgent antibiotic delivery from a baseline of around 30% to more than 80%.
One downside of any CQuIN lies in the mechanism itself – they do not result in extra funds flowing into hospitals, and therefore few invest heavily in their delivery. There is potential for convenience sampling of data fuelled by a Board-imposed demand to demonstrate success, which in some organisations may have resulted in processes aimed more at figures than true improvement.
In busy hospitals, and perhaps particularly in Emergency Departments, any quality improvement initiative for an individual condition risks diverting attention away from other conditions, and this valid concern has been levelled at the sepsis CQuIN. Coupled with Ambulance services’ adoption of care pathways using the NICE-revised Red Flag criteria, which have a tendency to over-alert in some age groups, this may have led to a ‘sepsis fatigue’ in some areas. We would counter that, given the limitations of available mechanisms to drive improvement, it was entirely appropriate to apply a CQuIN to a life-threatening and prevalent condition for which the previous standard of care ensured that only one in three patients were treated adequately.
The CQuIN undoubtedly saved lives, but we would not claim entirely without adverse consequence.
What about the biggest potential adverse consequence? Sepsis is almost unique among conditions in that improvement in its care likely helps individual patients, but at the potential expense of population-level harm. Antimicrobial resistance continues to present an existential threat to humans, yet sepsis improvement demands the use of antimicrobials.
This perceived adverse consequence did not seem to bear out in practice, fortunately. NHS England’s own data showed that antibiotic use in Emergency Departments did increase significantly during the period of the CQuIN, but total hospital consumption remained largely static. It would appear that the effect of the CQuIN was to ‘front load’ antibiotic delivery, but without tangible adverse consequence of increasing total use.
We submitted the below response to the Lancet in early November 2019, well within their publication deadline for response to correspondence. To date, the journal has elected not to publish (without satisfactory explanation): we provide it here for information.
Sepsis and antibiotic mindfulness: an alternative to hype and hyperbole
Nutbeam T, Daniels R
“Antibiotic use in emergency departments in English hospitals has doubled since 2015, coinciding with the introduction of the Commissioning for Quality and Innovation quality improvement initiative mandating antibiotic prescription within 1 hour of presentation” wrote Singer et al in their letter to the Lancet in October 20191.
Has the social media “hype” around this letter generated a potentially distorted picture of antibiotic use in English Emergency Departments (EDs), leaving those who aim to achieve the goals of the CQuIN or follow NICE or UKST guidance in fear of criticism, retribution and penalty? It is important to give context to the claims of Singer et al – to help with this we have reproduced the data provided to the original authors ad verbatim below (Table 1, Howard P, Rx-Info Define, personal communication).
This table disputes the relationship between the introduction of the CQuIN and increased antibiotic use in English EDs. The lowest rate (6%) of growth in antibiotic use in the last 9 years was in 2016/17 – post CQuIN introduction. Furthermore, the data reports DDD (Defined Daily Dose) – it does not report the number of patients commenced on antibiotics. NHS Digital’s Hospital Accident and Emergency Activity Data for the same period2 shows a steady growth in patients staying in EDs for 12 hours of more, with the total growth since 2014/15 in excess of the growth in DDDs of antibiotics – patients who are likely, if antibiotics have been commenced, to receive more than a single dose before discharge from ED. Yes, more antibiotics are being used in the EDs of English hospitals, but isn’t this just as likely a function of change of location of delivery of antimicrobials (due to ED service pressures, exit block and crowding) as opposed to a significant increase in their total use?
If more patients were started on antimicrobials in EDs, we would expect total hospital consumption to increase. It has, but barely: this alternative interpretation of the original data is entirely in keeping with the fact that total antibiotic use per 1,000 admissions to English hospitals has increased by only 1% since the introduction of the CQuIN. Singer et al were provided with the data in Table 1 yet elected not to cover this important context in their letter. Is this evidence of “sepsis hysteria”… maybe?
Mindfulness of the synergies, as well as perceived conflicts between messaging regarding sepsis and antimicrobial resistance, is surely the cure for this unhelpful hysteria. Balance in this domain has always been a challenge. Considerations must be applied for each and every patient unwell with infection including: beneficial early antibiotic therapy3, delayed antimicrobial decision-making, decision-making around appropriateness of care and of course alternative diagnoses. It was precisely with this is mind that, in mid-2019 and well in advance of Mervyn Singer’s letter, the UKST updated the widely-used4 “Sepsis Six” to prioritise senior clinical involvement (now Step 1 of the Six)5. Timely senior involvement with all these patients, supporting informed and balanced antimicrobial decision-making is in the best interests of our patients and our valuable antibiotics – in short, we need to empower clinicians to deliver antibiotics only when they are required. However, in an increasingly creaky NHS, when immediate senior review is not always available, we need redundancy, empowerment to act in time-critical situations and safety nets – this is where tools such as the UKST’s Red Flag Sepsis recognition tool and Sepsis Six treatment pathway play their valuable part.
We welcome, support and agree with many of the themes of Singer et al in their letter – so why this reply? Simple. Hysteria is best not countered with further hysteria. Mindfulness of our patients, our antimicrobials, our data interpretations and our conclusions must be the way forward. Working together, we can ensure that the data we need to guide decisions in this area are routinely collected, reported and used to influence our (mindful) decisions, giving our patients and our antibiotics the best deal they can get.
1) Singer M, Inada-Kim M, Shankar-Hari M. Sepsis hysteria: excess hype and unrealistic
expectations. Lancet. 2019 Oct 26;394(10208):1513-1514. doi: 10.1016/S0140-
2) Hospital Accident & Emergency Activity 2018-19, NHS-Digital, https://digital.nhs.uk/dataand-
accessed 1st November 2019.
3) Johnston AN, Park J, Doi SA, et al. Effect of immediate administration of antibiotics in
patients with sepsis in tertiary care: a meta-analysis. Clin Ther 2017;39(1):190-202.e6.
4) 2019 Annual Review by the All-Party Parliamentary Group on Sepsis, http://sepsisappg.com/wp-content/uploads/2019/07/APPGsepsis2019-Annual-Report.pdf, accessed 3rd November 2019
5) The United Kingdom Sepsis Trust, The Sepsis Six, https://sepsistrust.org/professionalresources/
clinical/, accessed 1st November 2019