Lancet on a wider study..

https://www.sciencedaily.com/releases/2015/05/150520193831.htm

Seems the extremes of either aren't the major killers. Probably because at the extremes people modify their behaviour more, take less risks.
But when the temperature is not extreme, but still either hot or cold, people more likely to risks.
Also wonder about population distribution amongst cold and hot climates. I'd guess more people live in the cold, due to water being available. Wonder if that skews the most deaths.