imagesWe know that hypothermia in sepsis is associated with increased mortality but other than that we tend to see fever in sepsis as something bad. We tend to perceive sepsis patients as more sick the more the temperature is elevated. We then tend to treat that hyperthermia with paracetamol, ibuprofen or external cooling.

A large Swedish study in Crit Care Med suggests that increasing temperatures might actually be a good thing. Increasing body temperatures on sepsis recognition is associated with improved outcomes.

The study
A large Swedish multicenter cohort study of data compiled in a national sepsis register. The study included 2,225 adult patients who were admitted to one of 30 Swedish ICUs with a diagnosis of sepsis or septic shock between 2007 and 2015. 56% were male and median age was 68y. Overall mortality was 24,7% and median length of ICU-stay was 13 days. The primary outcome was in-hospital mortality. Quality treatment was defined as the patient having sepsis bundle implementation within one hour.

Screen Shot 2017-03-25 at 14.39.16Mortality was, as expected, highest in the hypothermic (<35°C) group at 36.3%. But then the authors could demonstrate an almost linear reverse association between hyperthermia and mortality in the interval from 35°C up to greater than 41°C.

On average, mortality decreased by almost 5% for every increase of C°. As did ICU length of stay. As for clinical signs on presentation to ED, body temperature was the strongest predictor of mortality followed by respiratory rate, heart rate, saturation and then blood pressure.

One would suspect patients with higher body temperatures would be detected earlier and receive better treatment as more of them would have received bundle treatment within the first hour. However, sub-group analysis could demonstrate that the relationship between body temperature and survival remained intact whether stratified by age, lactate level, bacterial etiology, or sepsis bundle achievement.

Take-home message
There was a stronger and and inversely linear association between increasing temperatures at ED arrival and mortality as well as length of stay. Better quality of care in febrile patients did not explain this association. Of the vital signs, temperature was the best predictor of death.

In  an editorial in the same issue,  the editors Laupland and Niven write  that:

¨Although we must continue to use fever as a sentinel sign for detecting patients with serious infections, it is the absence of that sign that calls for our greatest attention among patients with severe sepsis and septic shock.¨

The paper lives here:

Crit Care Med. 2017 Apr;45(4):591-599.  Fever in the Emergency Department Predicts Survival of Patients With Severe Sepsis and Septic Shock Admitted to the ICU. Sundén-Cullberg J1, Rylance R, Svefors J, Norrby-Teglund A, Björk J, Inghammar M.

This entry was posted in Intensive Care. Bookmark the permalink.


  1. Pingback: Is Fever the New Hotness in Sepsis? - R.E.B.E.L. EM - Emergency Medicine Blog

  2. Pingback: Sepsis – feber – Mind palace of an ER doc

  3. Gary Gaddis says:

    I have long wondered why we clinical types would even think of adding antipyretics to interrupt a natural process, when patients are febrile. Exceptions could be if fever causes elevated cardiac output in someone with prior heart failure. Or a person is sufficiently miserable. Or to see if a somewhat listless child seems to be acting normally if/when fever is ameliorated.

    Other than that, why treat a fever?

    I never give myself antipyretics for fever. So, for me and my immune system, when under infectious attack, it’s like the rioters said in the 1960s during the conflagrations of the summer of 1968: Burn, baby. Burn!

  4. Pingback: LITFL Review 276 • LITFL • Life in the Fast Lane Medical Blog

  5. Pingback: SGEM#195: Some Like It Hot – ED Temperature and ICU Survival | The Skeptics Guide to Emergency Medicine

Leave a Reply

Your email address will not be published. Required fields are marked *