Objectives: To evaluate knowledge translation after publication of the target temperature management 33°C versus 36°C after out-of-hospital cardiac arrest trial and associated patient outcomes. Our primary hypothesis was that target temperature management at 36°C was rapidly adopted in Australian and New Zealand ICUs. Secondary hypotheses were that temporal reductions in mortality would be seen and would have accelerated after publication of the target temperature management trial.
Design: Retrospective cohort study (January 2005 to December 2016).
Setting: The Australian and New Zealand Intensive Care Society Centre for Outcome and Resource Evaluation adult patient database containing greater than 2 million admission episodes from 186 Australian and New Zealand ICUs.
Patients: Sixteen-thousand two-hundred fifty-two adults from 140 hospitals admitted to ICU after out-of-hospital cardiac arrest.
Interventions: The primary exposure of interest was admission before versus after publication of the target temperature management trial.
Measurements and main results: The primary outcome variable to evaluate changes in temperature management was lowest temperature in the first 24 hours in ICU. The primary clinical outcome variable of interest was inhospital mortality. Secondary outcomes included proportion of patients with fever in the first 24 hours in ICU. Mean ± SD lowest temperature in the first 24 hours in ICU in pre- and posttarget temperature management trial patients was 33.80 ± 1.71°C and 34.70 ± 1.39°C, respectively (absolute difference, 0.98°C [99% CI, 0.89-1.06°C]). Inhospital mortality rate decreased by 1.3 (99% CI, -1.8 to -0.9) percentage points per year from January 2005 until December 2013 and increased by 0.6 (99% CI, -1.4 to 2.6) percentage points per year from January 2014 until December 2016 (change in slope 1.9 percentage points per year [99% CI, -0.6 to 4.4]). Fever occurred in 568 (12.8%) of 4,450 pretarget temperature management trial patients and 853 (16.5%) of 5,184 posttarget temperature management trial patients (odds ratio, 1.35 [99% CI, 1.16-1.57]).
Conclusions: The average lowest temperature of postcardiac arrest patients in the first 24 hours in ICU rose after publication of the target temperature management trial. This change was associated with an increased frequency of fever not seen in the target temperature management trial.