What's new in critical illness and injury science? The use of risk stratification tools in patients with suspected sepsis in the acute care settings

Timely identification of sepsis remains a challenge, particularly in the emergent and acute care settings. This may be due to varied patient presentations, shifting diagnostic criteria, and challenges with applying existing diagnostic criteria in when incomplete information is available. This may be further exacerbated in the resource-limited settings. For these (and other) reasons, investigators have developed clinical scoring systems based on triage assessments and more readily available clinical criteria with the aim of aiding more rapid identification of patients with sepsis, and those at risk for clinical deterioration, including the need for intensive care unit admission and inpatient mortality. For many, confusion exists regarding the derivation, validation, and intended use of individual scores, leading to uncertainty as to which score is optimal for use in a given clinical setting. Furthermore, some researchers mistakenly ascribe diagnostic or prognostic attributes to scores and deploy their use in clinical studies in a manner that is not consistent with their intended use. Moreover, scores developed in the intensive care settings may overestimate mortality in the emergency department (ED).[¹]

One great limitation of sepsis literature is the varied means of diagnosing and identifying patients for study inclusion, which may introduce selection bias and result in the comparison of dissimilar populations within and between studies. This may range from studies that rely on sepsis (or sepsis related) diagnostic or billing codes without verification that the included subjects meet any (or even the same) diagnostic sepsis criterion,[²³] to those that restrict inclusion to a single set of diagnostic criteria (that has varied over time). This both impacts individual studies, hampers inter-study comparisons, and limits metanalyses, as populations defined by the sepsis-2, sepsis-3, or other criteria may be sufficiently dissimilar to impact results.

For example, the sepsis-2 criteria were rooted in the systemic inflammatory response syndrome (SIRS) criteria which, although sensitive for identifying sepsis, performed with a low specificity. This resulted in many individuals being identified as having sepsis (including study inclusion) who, in reality, did not have sepsis. Conversely, the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) significantly changed the definition of sepsis,[⁴] defining it as life-threatening organ dysfunction caused by a dysregulated host response to infection. SIRS was abandoned, and sepsis was now recognized by a change in sequential organ failure assessment (SOFA) score of ≥ 2. One consequence of this shift in the diagnostic criteria was a potential delay in recognition until organ dysfunction is readily established and observable. This appeared potentially contradictory to the longstanding mantra of treating sepsis quickly and aggressively. Moreover, the use of SOFA in ED and ward patients may be limited as some variables may not be promptly available or indicated in the acute care setting. Owing to this limitation, sepsis-3 also introduced quick SOFA (qSOFA) as a bedside tool “for the rapid identification of patients who are more likely to have poor prognosis.” Although qSOFA may be valuable in predicting sepsis-related mortality, it performed poorly as a screening tool for identifying sepsis in the ED and relying on qSOFA alone may delay the initiation of evidence-based interventions known to improve sepsis-related outcomes.[⁵] As such, researchers have both developed and assessed new and preexisting risk stratification tools to aid the rapid identification and prognostication for suspected sepsis cases. Table 1 summarizes some of these clinical tools, information on the derivation cohorts, intended use, and the types of criteria incorporated.

Table 1: A comparison of common risk stratification tools

Given the deployment of stratification tools in varied (or even unintended) clinical populations, the use of varied threshold values for scores and shifting diagnostic criteria, it should not be surprising that opinions vary greatly regarding which (if any) score should be deployed in the clinical settings. Among the most promising tools that outperform qSOFA are early warning scores such as the national early warning score (NEWS), the revised NEWS-2, and sepsis-focused tools such as the mortality in ED sepsis (MEDS) score[⁶] and the risk-stratification of ED suspected sepsis (REDS) score.[⁷⁸⁹]

NEWS was developed in the 1990s with the purpose of detecting patients who were deteriorating and was introduced nationwide in the United Kingdom in 2012.[¹⁰] It contains similar parameters to qSOFA but with the addition of other readily available triage parameters, including oxygen saturations, level of supplemental oxygen, temperature, and heart rate.[¹⁰] While not designed specifically for sepsis, NEWS (and later NEWS2) has proven effective in identifying patients with sepsis at risk of poor outcomes, and a positive score should trigger clinicians to investigate further for organ dysfunction (SOFA score) or consider escalation of therapy.[¹⁰] One advantage to NEWS2 (over MEDS or REDS) is that it does not require the diagnostic or laboratory testing and may be applied at triage.

In the current issue of the International Journal of Critical Illness and Injury Science, Verma et al. again report improved the sensitivity of NEWS2 compared to qSOFA for identifying septic patients.[¹¹] As this is known and well established, we encourage future efforts to focus more on comparing higher performing scores such as NEWS2, MEDS, and REDS to greater benefit those in the clinical setting. The availability of component variables at triage makes NEWS2 an attractive candidate for such use; however, a clearer understanding of the sensitivity and specificity of NEWS2 as compared to MEDS, REDS, and others is needed.

Research quality and ethics statement

This report was exempt from the requirement of approval by the Institutional Review Board/Ethics Committee. The authors followed applicable EQUATOR network (http://www.equator-network.org/) guidelines; however, no specific guideline is available for the editorials.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.