COMMENT This report from the TOBY trial is an important contribution to the literature on a salutary effect of hypothermia on outcomes in infants with perinatal asphyxial encephalopathy. These are among the sickest children cared for in anNICU and include themost severe adverse neurodevelopmental outcomes. Of the 3 large trials published to date (CoolCap, NICHD, and TOBY), two (including this one) did not achieve a statistical reduction in the primary outcome of death or severe neurological outcome. However, the relative risks in all three studies are actually very similar, ranging from 0.72 to 0.86. This suggests that there is actually a lowering of the risk of death or severe neurologic disability of about 25%; the lack of statistical significance may derive from the fact that the studies were powered to detect somewhat larger differences (although the current study enrolled 20% more patients than needed by their initial power analysis). Meta-analysis combining the various results does indeed suggest that hypothermia is beneficial to selected infants with severe perinatal asphyxia. And therein lies the rub. Hypothermia has very rapidly become the standard of care, so future randomized trials will be difficult to organize. Yet we do not know precisely which infants benefit the most (only those with severe asphyxia or those with only moderately severe entry criteria), which method (selective head cooling or total body cooling) works best, or even what is the optimal cooling temperature and duration of therapy. Nowadays, I would not in good conscience withhold hypothermia in any severely asphyxiated infant, yet I am troubled by the very basic clinical questions that remain. One interesting finding in this study was the fact that almost 25% of the noncooled infants had rectal temperatures of >381C (100.41F). Is it hypothermia or prevention of hyperthermia that is most important? The TOBY study also found improvement in a variety of secondary neurological outcomes, including,most importantly, the percentage of survivors with no severe neurodevelopmental deficits. Cerebral palsy and gross motor abnormalities were also less common in the cooled group, and Bayley Mental Developmental Index and Psychomotor Developmental Index scores trended higher. While these are gratifying results, it is very important to note (as the authors do in their summary), that neurodevelopmental testing at 18 months, as well as developmental testing later in childhood could affect the results, and will be necessary before the final assessment of this new and exciting therapy can be written.