The big thing is that the Nazis would have been perfectly "OK" with repeating the experiments as many times as needed. They kinda killed 6 million people, because they wanted them gone. They had absolutely no problems with how that killing was done; Killing the people was a
desired outcome. Because it was a
desired outcome, they did not take much care at all on other variables in their "studies". Such as subject health, fitness, mental condition, etc. Likewise there was no attempt at diverse subject populations (because it was ethnic cleansing!), so any anomalous results that might be present in a wider human population were not given any attempt at being tested for, etc. The studies were just plain shit. If you want a reason not to use this data, this is the reason you should reach for first. They were poorly constructed,
BECAUSE they sought to kill the subject, and made no efforts to control for the host of biological processes and effects they were introducing into their experiments through such systemic maltreatment. Because they were conducted this way, they are garbage. End of story.
The philosophical question of "taint of evil" in anything that comes from that research is moot; All subsequent studies were informed by what was learned in that study, and that knowledge was used to STRUCTURE the more ethical, better designed, and more informative trials, which give more modern (and reproducible) data.
As a consequence, if we ascribe to the "taint" philosophy, *ALL* hypothermia data, (unless you can prove that your researchers have lived under a proverbial log, and learned everything independently) from all later trials are contaminated, no matter how ethically they were performed.
It's an insane position to take.
Don't use the nazi studies because they were poorly controlled, and irreproducible. There are better datasets that are less ethically dubious and of higher quality. Use those instead.
But taint of evil? No, that's bullshit. It's about process,
not magical hoodoo.