Some of the above seems to paint a fairly idyllic picture of the life of animals in research. When considering the suffering of these animals, I think it is important to keep in mind that they are often “recycled” repeatedly from one experiment to the next and from institution to institution. Thus, a given animal’s life may consist of years of suffering different types of experiments along with suffering associated with transport between institutions, often over great distances. All of this is followed by a death that may include suffering of its own.
It is also important to consider that many of the practices used on farmed animals were first tested on animals in research settings who suffered horribly as a result. Using the ventilation shutdown example cited above, many hens suffered gruesome deaths filled with prolonged suffering while the agriculture industry tested methods that would later be productionized on farms. Granted, the numbers of animals were fewer than when put into practice, but we shouldn’t forget that animals used in research also includes farmed animals.
You can look at APHIS inspection reports and see many incidents of animals dying for reasons other than euthanasia. A better source is NIH OLAW reports that you can obtain via FOIA. There’s no shortage of research institute self-reporting of animals dying due to a variety of issues including starvation, dehydration, improper protocol, etc.