>>Savella
You could make a virtually identical case against Prozac or any of the other marketed SSRI's. All of them needed repeated trials to come up with two trials showing efficacy, and a lot of the safety issues (except for the hypertension which I agree is somewhat worrying) overlap as well.
My take on the less-than-stellar demonstrated efficacy issues with all these psychoactive drugs is that it's always going to be hard to demonstrate efficacy on a fuzzy endpoint with high variability. Further, I believe different people respond to different drugs, and so just because in an overall trial efficacy is only moderate, doesn't mean that you won't get significant efficacy in particular patients. (Same is true with say migraine prophylaxis).
Peter