Interpreting the results of the targeted temperature management trial in cardiac arrest.

The targeted temperature management (TTM) trial, which found that cooling to 33°C after witnessed cardiac arrest (CA) conferred no benefits compared with 36°C, has led to much debate in the hypothermia community. This article discusses what lessons can be drawn. The TTM trial achieved far better outcomes in controls than any previous randomized controlled trial (RCT) or any nonrandomized study where no fever control was applied. On the other hand, rates of good outcomes in the hypothermia group were somewhat lower than in previous RCTs and most nonrandomized studies. The TTM authors conclude that benefits of temperature management are derived exclusively from fever control and that further lowering of temperature confers no benefit. Indeed, without doubt, the TTM trial demonstrates the crucial importance of strict fever control after CA and that this provides sufficient neuroprotection for some patients. However, we argue that the hypothermia intervention was executed suboptimally (possibly inadvertent selection bias; late start of cooling, up to 4 hours after ROSC; slow cooling rates, 10 hours to target temperature; more rapid rewarming than previous studies; and some other issues). This could explain high rates of good outcomes in controls and lower-than-expected rates in patients cooled to 33°C compared with previous randomized and nonrandomized studies. Outside of two previous RCTs, the use of hypothermia after CA is supported by hundreds of animal experiments, evidence from 46 before-after studies and large registries, and indirect supporting evidence from 7 RCTs in newborns with neonatal asphyxia. In addition, one RCT found improved outcomes with 32°C compared with 34°C. It remains to be explained why the TTM results so completely contradict previous studies in this field. These issues should be thoroughly discussed before changes in guidelines and protocols are made. Ending or modifying hypothermia treatment after CA should require the strongest possible evidence.