Drug Testing In Mice May Be a Waste of Time, Researchers Warn 148
An anonymous reader writes "A group of researchers including Dr. H. Shaw Warren of Mass. General Hospital and Stanford genomics researcher Ronald W. Davis have published a paper challenging the effectiveness of the 'mouse model' as a basis for medical research, based on a decade-long study involving 39 doctors and scientists across the country. In clinical studies of sepsis (a severe inflammatory disorder caused by the immune system's abnormal response to a pathogen), trauma, and burns, the researchers found that certain drugs triggered completely different genetic responses in mice compared with humans. The Warren-Davis paper was rejected by both Science and Nature before its acceptance by the Proceedings of the National Academy of Science, perhaps suggesting the degree to which the 'mouse model' has become entrenched within the medical research community. Ninety five percent of the laboratory animals used in research are mice or rats. Mice in particular are ideal subjects for research: they are cheap to obtain and house, easy to handle, and share at least 80 percent of their genes with humans (by some reckoning, closer to 99 percent). Over the past twenty five years, powerful methods of genetically engineering mice by 'knocking out' individual genes have become widely adopted, so that use of mice for drug testing prior to human clinical trials has become standard procedure."
Peer review (Score:0, Interesting)
Being rejected by Science and Nature might also be indicative of being bad science. Not reading the report yet, the options seem to be intellectual dishonesty from some of the most respected sources of science, or the mice findings are fundamentally flawed. On the outset, I think being rejected by big names in science is usually pretty telling.
Bacteria as a major clue (Score:5, Interesting)
Yet there was always one major clue that mice might not really mimic humans in this regard: it is very hard to kill a mouse with a bacterial infection. Mice need a million times more bacteria in their blood than what would kill a person.
“Mice can eat garbage and food that is lying around and is rotten,” Dr. Davis said. “Humans can’t do that. We are too sensitive.”
I agree but there are reasons why we use mice (Score:3, Interesting)
1. Mice have no lobby.
2. Mice have shorter lifespans.
3. You freak out every time we use chimps or human analogues in the simian world.
4. Mice are easier to squish between plates to measure changes, especially when we use flourescent tags on the meds or target we're looking at, so we don't have to cut them up to find out what's going on.
(yes, my point 4 is really what happens - we used to cut them up before we figured out how to make them glow with jellyfish gene tags - and once you cut open the brain, it's game over)
Re:Rejection or Science Nature (Score:5, Interesting)
Its very hard to publish there, but the quality of publications is not that high, possibly even lower than elsewhere if you measure by false positive rate. There is a mass failure to understand the importance of the assumptions underlying statistical inference (as you mentioned), as well as the importance of completely reporting your methods and data so that it is possible for others to intelligently draw their own inferences and replicate your work. In short, those journals have a culture that encourages "sexy" and "conclusive" results at the expense of the fundamental basis for successful science that we learn in gradeschool.