None of those are inherently fatal, either alone or in combination. To illustrate, imagine I did a study where I took 20 people, subjected them to a 10 minute ice bath, followed by a chest X-ray, and then put them naked in a steel cage and lowered the cage to the bottom of a swimming pool for 20 minutes (yes, the pool is full of water). Study participants are not told beforehand that they will be going into water.

And suppose further that 9 of them are still alive when I pull them out of the water. Not only are they alive, they are conscious, alert, and show no ill effects from 20 minutes underwater.

That would be a study worth talking about despite the sample size and lack of control and short time span.

Those would not be a fatal problems because we already know a lot about what happens to humans in general who spend 20 minutes without diving equipment underwater. We have a good idea of what the probability is for a human to survive such a thing, and can estimate the probability that my sample included 9 such people. It would be close to zero. Heck, even if I had the misfortune of accidentally getting my sample at a free diving competition the chances that I'd get 9 people who could hold their breaths for 20 minutes would be very low.