The Normal Abnormal Vital Sign of Sepsis*
In this issue of Critical Care Medicine, Sundén-Cullberg et al (3) report their investigation of the role of body temperature and outcome among patients with severe sepsis and septic shock. They studied 2,225 adult patients admitted to 30 ICUs in Sweden between 2007 and 2015. Temperature was documented at presentation to the emergency department and subsequently explored in relation to all-cause in-hospital mortality. They observed that fewer than one half of patients had an elevated body temperature at emergency department admission, yet elevated temperature was associated with a decrease in-hospital mortality. After controlling mortality analyses for several potential covariates not limited to age, sex, underlying comorbidity, and time to receipt of antibiotics, elevated body temperature at presentation was consistently associated with improved outcome.
This study adds to a growing body of literature investigating the effects of temperature on outcomes of patients admitted to the ICU (2, 4–9). The impact of temperature abnormalities among ICU patients and outcome is a complex issue. Although the results have varied among patient subgroups, large cohorts in North America, Asia, Oceania, and Europe have generally shown that fever either has no effect or reduces mortality in most patients with infections and that hypothermia is associated with adverse outcome (2, 4–9). Interventional studies have shown conflicting results with temperature management strategies in critically ill patients (10–13).
An important and novel aspect of the current study is that unlike previous investigations that evaluated temperatures following ICU admission, the investigators documented the first temperature obtained in the emergency department. Although a seemingly minor point, it is of importance as this pretreatment measurement removes the clouding influence of potential (non)-responders to treatment that may occur by time of ICU admission. The investigators show that increased temperature is associated with improved outcome, and this analysis was robust to inclusion of numerous confounding variables. As detailed by the authors, a number of limitations of the study must be kept in mind not limited to lack of standardized temperature measurements and lack of severity of disease assessment. Furthermore, it tells us nothing about whether therapeutic temperature manipulations may have influenced outcome.
This study raises several important questions for further consideration. It is of particular interest to know what the determinants of temperature at presentation are. The infecting organism or focus may be important, and there may also be aspects unrelated to the infection per se including environmental exposures, comorbid illnesses, medications, and intoxicants. Host factors and immune responsiveness determinants could also play a role. One may also speculate that patients presenting with normal or lower temperature may be further along the sepsis pathway or delayed in their presentation. Although the present study is valuable to identify the role of temperature, answers to these questions must be sought in other investigations.
This study implies that fever may be the “normal” temperature of severe sepsis and septic shock. Clinicians would largely agree that detection of fever is a useful abnormal sign suggesting the possibility of infection in a patient.