Study populationThe study was performed at in 3 Emergency Departments and10 ICU’s in three large hospitals in Pittsburgh, PA (Presbyterian-Montefiore, Mercy, and Shadyside hospital). Consecutive patientswho remained comatose after OHCA or IHCA were screened for thestudy. The only exclusion criterion was a decision not to use TH,based on pre-existing conditions (e.g. terminal disease). All hospi-tals used the same protocol for treating post-CA patients, including24-h cooling to 32 or 33 ◦C followed by slow rewarming. Patientswith (high risk for) active bleeding were cooled to 35 ◦C; data forthese patients were collected, but not included in the compara-tive analyses because hypovolemic shock can affect temperature.The University of Pittsburgh IRB approved the study and waivedthe need for informed consent because no changes were made inclinical treatment.The target temperature was determined by the attending physi-cian before screening; according to our protocoltargettemperaturecould be 32, 33 or 34 ◦C(35 ◦Cif active bleeding waspresent).Choiceof cooling method was random or based on availability of cool-ing devices, unless contraindications for specific cooling methods(e.g. recent history of deep venous thrombosis, skin disease) werepresent.Cooling efficacy was measured as follows. In the induction phasethe aim was to reduce core temperature to target as rapidly as pos-sible; in the maintenance phase, to maintain temperature close totarget for 24 h; and in the re-warming phase, to achieve slow andcontrolled re-warming (maximum 0.25 ◦C/h) [14–15].