Sepsis, Critical Care, Critical Care Alert

Critical Care Alert: Prognostic Accuracy of Sepsis-3 Criteria for In-Hospital Mortality Among Patients with Suspected Infection

A 78-year-old female presents to the emergency department with altered mental status. Vital signs are notable for a temperature of 37.7 degrees Celsius, respiratory rate of 24 breaths per minute, heart rate of 89 beats per minute, and blood pressure of 104/68. Work-up is notable for a normal head CT, white blood cell count of 10,000, and urinalysis with nitrites and bacteria. Her infectious source was noted but she did not meet criteria for the definition of sepsis so antibiotics and fluids were not started until the patient was admitted to the floor. She subsequently becomes hypotensive requiring transfer to the Intensive Care Unit.


Freund Y, Lemachatti N, Krastinova E, et al. Prognostic Accuracy of Sepsis-3 Criteria for In-Hospital Mortality Among Patients with Suspected Infection Presenting to the Emergency Department. JAMA. 2017;317(3):301-308.


To prospectively and externally validate the quick Sequential Organ Failure Assessment (qSOFA) as a predictor of mortality and to compare the performance of qSOFA with systemic inflammatory response syndrome (SIRS) criteria.


Sepsis is one of the most common conditions responsible for admission to the Intensive Care Unit (ICU) and is associated with high in-hospital mortality. New definitions for sepsis and septic shock were introduced in 2016 by the international Sepsis-3 task force. Sepsis is now defined as life-threatening organ dysfunction caused by a dysregulated host response to an infection. This definition replaces the previous definitions of sepsis which was based on SIRS criteria and suspicion for infection. “Severe sepsis,” or sepsis with lactic acid elevation, is no longer recognized as an entity. SOFA is an extensive scoring system designed for ICU patients that is not practical for use in the Emergency Department (ED). The Sepsis-3 task force validated the qSOFA as a surrogate for the SOFA for screening for sepsis, but ED populations were not specifically studied. The qSOFA gives points for respiratory rate >21 breaths per minute, systolic blood pressure  ‰¤100 mmHg, and altered mental status with a score of 2 being indicative of sepsis. Many emergency physicians have criticized Sepsis-3 because it did not specifically involve ED patients or physicians. This study aimed to compare the ability of the qSOFA to predict mortality compared to SIRS and lactic acid measurements in an ED patient population.


    • International, multicenter prospective cohort study involving 30 European centers

    • Enrolled consecutive adult patients over a 4-week period from May to June 2016 who visited one of the recruiting EDs with a possible infection

    • Utilized “worst” level of components of qSOFA and SIRS for data collection and collected as much of the SOFA data as was available

    • Altered mental status defined as Glasgow Coma Scale (GCS) or clinical judgement of the treating physician

    • Missing SOFA data was assumed to be within normal limits

    • Calculated sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) for a qSOFA  ‰¥2 and created receiver operating characteristic (ROC) curves for each set of sepsis criteria

Inclusion criteria: Adult patients presenting to the ED with clinical suspicion of infection

Exclusion criteria: Infection not confirmed, “low acuity” infections defined as localized infections with normal vital signs that did not require laboratory studies (tonsillitis, cystitis, skin abscess, etc.), pregnant women, prisoners

Primary outcome: In-hospital mortality within a 28-day period of admission

Secondary outcomes: Admission to the ICU, length of ICU stay greater than 72 hours, and a composite of death or ICU stay greater than 72 hours

Comparison of Endpoints


Comparison of Test Characteristics for In-hospital Mortality

*Severe sepsis = SIRS > 2 plus lactic > 2 mmol/L


Authors' Conclusions and Limitations


  • qSOFA performed better than of the other classification systems for predicting in-hospital mortality
  • This study allows for the beginning of a paradigm shift for classification of sepsis patients from the ED
  • SIRS is not specific enough to be a reliable predictor of outcomes for patients with presumed infections
  • Mortality rate of 3% in patients with qSOFA < 2 suggests that ill patients will not be missed
  • No added benefit of adding lactic levels to qSOFA
  • qSOFA can easily be assessed immediately upon a patient's arrival in the ED without blood work and therefore allows more rapid detection of potentially critically ill patients
  • Study limited by lack of follow-up for discharged patients, use of the worst value in calculation of the qSOFA, and missing data prevented accurate SOFA calculation
  • Additional limitations of low overall mortality rate, lack of blinding, and inclusion of patients who were “do not resuscitate”

Our Conclusions and Applications for the ED


  • qSOFA is an easy tool for ED physicians and nurses for rapid screening for sepsis and a concerted effort should be made to transition away from use of SIRS to define sepsis.
  • SIRS is likely responsible for over-diagnosis of sepsis because of its high sensitivity with very low specificity.
  • Fever, tachycardia, and leukocytosis should not be ignored by any means, but they should be emphasized less in regards to predicting a patient's outcomes.
  • Additionally, absence of SIRS criteria should not cause an ED provider to miss a critically ill septic patient in the setting of altered mental status or hypotension.
  • Establishing common definitions with internists and intensivists can allow for more effective hand-offs and more appropriate dispositions for patients.