Sepsis should be treated within one hour, says NICEBMJ 2017; 356 doi: https://doi.org/10.1136/bmj.j1257 (Published 10 March 2017) Cite this as: BMJ 2017;356:j1257
All rapid responses
Early identification of patients with sepsis saves lives. The recently announced UK National Institute for Health and Care Excellence (NICE) quality standards are therefore laudable in principle. The practical reality of applying them to the broader definition of sepsis outlined by the 2016 NICE sepsis guidance is unclear.
That a new definition was needed is in little doubt. There had long been an appreciation that the 2001 Systemic Inflammatory Response Syndrome (SIRS) based definition was inadequate being insufficiently sensitive, or indeed specific. NICE anticipated the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) and removed SIRS from its definition producing a more inclusive version based on broader clinical criteria (e.g. confusion, ashen appearance) with the intention of identifying people who would not previously have been categorized as septic.
Operationalizing these guidelines may therefore result in an abrupt increase in the number of people requiring urgent assessment – many perhaps unnecessarily. This will have implications for NHS hospitals already facing limited staffing, increased patient numbers, and a requirement for good antibiotic stewardship. In addition there may be financial penalties for failing to achieve government targets. A current NHS England “Quality and Innovation” (CQUIN) target mandates administration of antibiotics within 60 minutes for patients labeled “septic”.
To establish the proportion of patients classified as “septic” by NICE guidance compared to previous SIRS-based criteria, we retrospectively reviewed the initial medical clerking of 415 consecutive patients presenting over seven days to the acute medical unit of our urban general hospital. The clinical information necessary for the application of the two sepsis definitions was collected. An infection specialist independently reviewed the records and indicated whether they felt the patient required urgent treatment for sepsis. In the absence of any other, this was used as the gold standard against which to compare the various screening scores (data submitted).
SIRS-based criteria classified 20% (95% CI 16 – 24) as septic whereas NICE-recommended criteria identified 33%, (95% CI 29 – 38). The Sepsis-3 proposed “qSOFA” screening tool defined just 4% as septic.
Compared to a consultant’s clinical impression of sepsis, SIRS-based criteria had a sensitivity of 76% (95% CI 60-87) and specificity of 87% (95% CI 83 – 90). Using NICE sepsis guidelines sensitivity was similar at 84% (95% CI 71-94) but specificity fell to 73% (95% CI 47-60). qSOFA scoring had the lowest sensitivity at 13% (95% CI 5 – 27) but highest specificity at 95% (95% CI 95 – 99).
22 people were admitted to ITU or died within 14 days. Of these 9 were identified by both SIRS-based and NICE sepsis definitions. No additional patients were recognised by SIRS-based criteria alone but NICE guidance identified 7 additional patients. qSOFA identified 7 patients from the death/ITU group all of which were identified by NICE and 5 by SIRS-based definitions.
We concluded that implementing NICE sepsis guidance will have an operational impact with one third of acute medical take being classed as requiring urgent antibiotic administration. It is not clear to what extent this is clinically justified - improved sensitivity has come at the cost of significantly reduced specificity.
Given the staffing pressures and limited extent of immediate senior medical availability, particularly overnight, it is likely that this will lead to an increase in the administration of unnecessary – or at least unnecessarily broad-spectrum – antibiotics with implications for stewardship, complications and the risk of later bacterial resistance. We are aware that junior doctors – those making the point of care decisions – tend to associate the label of “sepsis” with the need for broader-spectrum antibiotics, often inappropriately. In our cohort 54% of patients labeled as septic by NICE had some form of chest infection and would have been well managed via standard chest infection management pathways. In addition the quality standards, combined with well-intentioned government CQUIN target for sepsis, may have unintended consequences: it has become apparent that patients are being given antibiotics hours before they are formally assessed. In part this may be the face of good medicine in the context of pressurized system – but hospitals need to be cautious that it does not become primarily a means to avoid a financial penalty. Contemporary meta-analyses of studies of antibiotic timing are not universal in their support for the one hour target – the benefit is more likely to be detectable in cohorts of particularly high mortality. Early antibiotics are after all just one component of the management of those with infections and whilst a CQUIN for antibiotic administration is easy to measure there are many other unmeasured aspects of care that may be more important.
Sepsis-3 very deliberately did away with the concept of severe sepsis, implying that all sepsis was by definition, life-threatening. However a tool that classifies 33% of medical admissions as septic needs further refining, perhaps even a return to risk stratification by incorporating a score such as qSOFA. Whilst not sensitive enough to be used alone, qSOFA does appear to be very effective at recognizing those patients likely to experience an adverse outcome and therefore most likely to benefit from urgent senior medical attention.
In conclusion, hospitals need to implement NICE guidance cautiously and in parallel with robust antibiotic stewardship systems. NICE recommendations need to be subject to ongoing prospective research with the aim of clarifying their sensitivity and specificity and identifying how they can be improved.
Dept of Infection and Dept of Acute Medicine,
Heart of England NHS Foundation Trust, Birmingham.
1 Suspected sepsis: summary of NICE guidance. 2016;354:i4030.http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id...
2 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. 2003. 1250–6. doi:10.1097/01.CCM.0000050454.01978.3B
3 The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). 2016. 801–10. doi:10.1001/jama.2016.0287
4 Sterling SA, Miller WR, Pryor J, et al. The Impact of Timing of Antibiotics on Outcomes in Severe Sepsis and Septic Shock: A Systematic Review and Meta-Analysis. Crit Care Med 2015;43:1907–15. doi:10.1097/CCM.0000000000001142
5 Donnelly JP, Safford MM, Shapiro NI, et al. Application of the Third International Consensus Definitions for Sepsis (Sepsis-3) Classification: a retrospective population-based cohort study. Lancet Infect Dis 2017;0. doi:10.1016/S1473-3099(17)30117-2
Competing interests: No competing interests