Thread by @HengartnerMP

Obviously Jauhar and Hayes do not want to end the „war on antidepressants“, as now they even made text changes to the published version of their critique of the withdrawal review to defend their misleading statements. So, sadly enough, the debate has to continue… Thread

1/n
Here the critique they originally published: "[Davies & Read] seem also to misunderstand simple principles that underpin why blinded RCTs are necessary. Thus, in the trial by Montgomery et al. (2005), DESS score was higher during placebo treatment than active treatment.”

2/n
In response, Davies, Read and I pointed out that this is misleading and misrepresents the data, because this was a discontinuation trial, so more withdrawal symptoms in the group where the drug was replaced by placebo means that withdrawal occurred madintheuk.com/2019/02/end-th…

3/n
On Twitter, Jauhar then claimed that he merely quoted the wrong reference. In response I pointed out that even when he refers to Montgomery et al 2004, their sentence is misleading. Clearly I was right, because now they also changed the entire meaning of the sentence...

4/n
Inconceivably, Jauhar and Hayes changed the very meaning of this crucial sentence during proof reading in a paper published online. Here is the new text:
"Thus, in the trial by Montgomery et al. (2004), DESS score was numerically higher during active treatment than placebo”

5/n
So, it's not only that the reference was replaced, instead of “DESS score higher during placebo treatment” in the revised version it now reads “DESS score numerically higher during active treatment”. Of course an author is not allowed to do this during proof reading!

6/n
Nevertheless, and even worse, although they revised both reference AND text post-hoc, Jauhar and Hayes are again wrong! It’s incredible! So, here is the correct interpretation of Montgomery et al 2004, because its clearly different than stated by Jauhar and Hayes...

7/n
In the paroxetine arm there was a SIGNIFICANTLY HIGHER dess score when the drug was replaced with placebo (which indicates withdrawal), whereas in the agomelatine arm there was NO CHANGE (ie no withdrawal reaction)! This is the correct interpretation!

8/n
Here is the conclusion by Montgomery et al (2004): "By contrast to paroxetine, abrupt cessation of agomelatine is not associated with discontinuation symptoms".
See any mention that withdrawal was higher during active treatment? ncbi.nlm.nih.gov/pubmed/15289700

9/n
This whole debate makes me sad and angry. Changing the meaning of a crucial sentence in a published paper is not allowed! This is very bad behavior. Proof reading is only to correct typos and the like. I have never seen something like this before in my 10 years in research