Think sepsis: SIRS negative ≠ not septic

Another of the important critical care articles released by NEJM this week is this analysis on the sensitivity of the SIRS criteria for detecting severe sepsis.


The SIRS criteria

The  systemic inflammatory response syndrome (SIRS) criteria were described as markers of significant immune response, and are used in the setting of infection to identify the presence of sepsis. SIRS is traditionally diagnosed when two or more of these criteria are present:

  1. Temperature < 36°C or > 38°C
  2. Heart rate > 90 bpm
  3. Tachypnoea > 20 breaths per minute, or an arterial pCO2 < 32 mmHg
  4. White cell count (WCC) < 4000 cells/mm³ (4 x 109 cells/L) or > 12,000 cells/mm³ (12 x 109 cells/L), or > 10% immature neutrophils (left-shift)

Some centres broaden the criteria to include other markers, such as altered mental state and hyperglycaemia in the absence of diabetes.

Diagnosis of Sepsis

The SIRS criteria, in the setting of sepsis, are traditionally applied thus:

  • Infection + SIRS = Sepsis
  • Sepsis + hypotension + end-organ damage = Severe sepsis
  • Severe sepsis + refractory hypotension = Septic shock

On to the Study

This is a retrospective database review over 14 years. Over 1.1 million admissions to 172 ICUs across Australia and New Zealand were reviewed, yielding 109,663 admissions with severe sepsis—infection with organ dysfunction. Of these 87.9% were SIRS positive, leaving 12.1% SIRS negative, using the traditional definition of 2 criteria present (measured within the first 24 hours). The mortality rate in both groups decreased similarly over time, with the SIRS-negative group having a lower mortality (8.5% in 2013 vs 27.7% in the SIRS-positive group). Being SIRS positive significantly increased risk of death by 26% across time and place.


The most common SIRS criteria present were tachycardia, tachypnoea or low pCO2, followed by low WCC. There was a linear increase in mortality of 13% for each additional SIRS criteria met from 0 through to 4.


SIRS negative sepsis

Almost 1 in 8 septic patients were SIRS negative using the traditional definition. In 20%, no SIRS criteria were met at all. The most common single criterion met was raised WCC, followed by tachycardia.

The incidence, proportion, and mortality of SIRS-negative sepsis decreased over time almost identically to the rates among patients with SIRS-positive sepsis. The mortality rate in SIRS-negative patients was 27.7% (n = 361) in 2000 and decreased to 8.5% (n = 1315) in 2013. The quadrupling of admissions in this time period is not well explained—could this be better data collection, more ICU beds (or shorter admissions, allowing more admissions), sicker patients or better diagnosis of sepsis?


...the requirement of two or more SIRS criteria for the diagnosis of severe sepsis excluded a sizable group of patients in the ICU with infection and organ failure.

There was a linear rise in mortality with the presence of each additional SIRS criterion, with no transitional increase in risk with 2 criteria to justify this as a definitional threshold.


This was a retrospective study using data originally collected for quality control purposes and only includes the first 24 hours of ICU admission. It is possible that this would miss patients who met SIRS criteria before or after this period, or in between the data collection points. This may also miss patients who go on to develop organ failure after the first 24 hours.

Relevance in the ED

Despite being a study of ICU patients, there is some relevance for those treating septic patients in ED as well. To me, the key messages were:

  • SIRS negative ≠ not septic
  • SIRS criteria are neither specific (only ~1/3 SIRS positive patients have an infection) nor sensitive (as demonstrated in this study)
  • SIRS criteria evolve and may not be present early in the course of sepsis in ED
  • Temperature is the least sensitive criteria (abnormal in <60%)
  • Tachycardia is the most common
  • Being SIRS positive is still a marker of higher mortality in sepsis
  • More SIRS criteria = higher mortality



Share your thoughts