Temperature Variability in a Modern Targeted Temperature Management Trial

OBJECTIVES: The Eurotherm3235 trial showed that therapeutic hypothermia was deleterious in patients with raised intracranial pressure following traumatic brain injury. We sought to ascertain if increased temperature variability within the first 48 hours, or for 7 days post randomization, were modifiable risk factors associated with poorer outcome.

DESIGN: Eurotherm3235 was a multicenter randomized controlled trial. Patients were randomized to receive either therapeutic hypothermia in addition to standard care or the later only. Mean moving range (mr) was used to stratify subjects into tertiles by the variability present in their core temperature within the first 48 hours post randomization and within 7 days post randomization. The primary outcome measure was a collapsed Glasgow Outcome Scale-Extended at 6 months post randomization. The temperature variability effect was estimated with ordinal logistic regression adjusted for baseline covariates and treatment effect.

SETTING: Forty-seven critical care units in 18 countries.

PATIENTS: Patients enrolled in the Eurotherm3235 trial to either therapeutic hypothermia or control treatments only.

MEASUREMENTS AND MAIN RESULTS: Three hundred eighty-six patients were included in our study. High level of temperature variability during the first 48 hours was associated with poorer collapsed Glasgow Outcome Scale-Extended. This effect remained statistically significant when only the control arm of the study was analyzed. No statistically significant effect was seen within the first 48 hours in the hypothermia group or within 7 days in either group.

CONCLUSIONS: When targeting normothermia, temperature variability may be a statistically significant variable in an ordinal analysis adjusted for baseline covariates.