Posted by Elizabeth on August 15, 1999, at 1:32:51
I don't take kindly to being called "naive" just because I bring up professional ethics as though it were a real issue for some researchers. Please try to be more considerate. Thanks.
> David Michelson (lead author) works for Lilly.
A number of the other authors work for universities. So? I think before you accuse people of twisting their results around, you should have some evidence. If studies funded by other sources (say, NIMH) showed differently, then I'd be more convinced by accusations that Lilly fixed this one. As it is, I don't doubt their results. I think they are plausible.
> Also of note are the previous studies involving Prozac in which some of the authors have been involved. All that I have read were funded by Lilly. They also had results favorable to Lilly.
Post hoc, ergo propter hoc...?
> The study does not say there is no weight gain on Prozac. It says the weight gain on Prozac was equal to the placebo group weight gain.
Right; meaning that Prozac does not appear to "cause" weight gain any more often than an inactive placebo "causes" it. That is, people didn't gain weight on Prozac that they wouldn't have gained off it.
> On Prozac 25.4% of the participants had 7% or more increase in weight. In placebo it was 26.7%. The weight gain in the Prozac group was attributed to improvement in appetite after recovery while it was poor nutrition for the unfortunate placebo group.
Weight gain was associated with initial appetite loss. I don't see where it says that it was caused by poor nutrition in the placebo group. As far as I can tell, they attribute it to recovery from depression in *both* groups.
See, they only looked at people whose depression was remitted, whether they were on Prozac or placebo...that's why there were so many placebo dropouts by week 50, because most of the people on placebo relapsed. So they were comparing Prozac responders to placebo responders. IOW, they were controlling for the presence of depression. A study that compared Prozac to placebo without controlling for depression would probably make it appear that Prozac was causing weight gain because more of the placebo group would remain depressed (depressed people weigh less, as a whole, than nondepressed ones, though there are individuals who gain weight while depressed). That would be a design flaw.
> Their statement that "the number of fluoxetine-treated patients with a 7% or greater increase in weight never exceeded that of patients in the placebo group" loses some impact when you consider that only 15 placebo participants were able to finish the study.( It will be a great sound bite though) The results of all participants, whether they completed or not, were a little different.
> Prozac- 13.2% had 7% or greater weight increase.
> Placebo- 7.4% had 7% or greater weight increase.This isn't very impressive. The placebo dropouts had less time to gain weight because they left the study early; of *course* fewer of them are going to. You'd have to look at when they dropped out, and how much weight they gained per unit time, in order to make this meaningful.
Note that they also said: "Weight change did not vary among patients assigned to receive placebo who dropped out before completing 26 total weeks (12 weeks of acute treatment and 14 weeks of continuation therapy) compared with patients receiving placebo who completed 26 total weeks."
> I could go on and on ripping this apart but I think you get the picture. You can make numbers mean whatever you want.You haven't made them mean what you want, though. You haven't ripped anything apart. Adam's "I know my body" argument is far more convincing.
Look, all I'm saying is, this weight gain phenomenon doesn't seem to be well validated. Yes, it probably happens occasionally, but not enough to show up as statistically significant even in large trials. I think it's probably overreported by people who just fear the normal weight gain that comes with recovery. This is unfortunate indeed.
poster:Elizabeth
thread:10121
URL: http://www.dr-bob.org/babble/19990829/msgs/10121.html