Sameer Jauhar
@sameerjauhar
Clinical Reader in Affective Disorders and Psychosis Division of Psychiatry @ImperialBrains Views my own. Shanti.
Most viewed this week from @JAMAPsych: Systematic review and meta-analysis: Stopping antidepressants leads to slight symptom increases compared to placebo, but these symptoms are not clinically significant and do not include mood changes. ja.ma/3IGmK72
We disagree on basic points of methodology. You may also wish to discuss the antler trial with the authors. Interpret the increased symptoms, the fact that they're probably not normally distributed, within the bigger sphere. Examine how they reported their own findings
Survey data involves actual people - if some can have delayed onset clearly it is a property of humans. ANTLER (perhaps the word you have said most today) found maximal withdrawal symptoms at 4-8 weeks after stopping, not a survey. It was an RCT in the NEJM.
Quite a few mistakes in this piece. I don't think the authors actually read the paper. Best example is stating we excluded a long-term trial. It is in the appendix. Case closed, and my opinion. Retraction of the conversation or correction, I will let them decide.
When denial of the issue is no longer tenable they turn to minimisation and underplaying. See @ConversationUK article here: bit.ly/4eFrZzO (31/n)
Citing.....survey data. This is why we put together the evidence.
They also ignore delayed onset withdrawal effects. 1 study found average time to onset of withdrawal effects was 4 weeks (SD 13 weeks) bit.ly/3GAQPnW, matching the ANTLER trial which found maximal withdrawal effects 4-8 weeks after stopping bit.ly/3Gyty6a (23/n)
Eh? No data informed analysis here. Did this gentleman ? WHO read the paper or evidence or base thinking on X?
Thank you so much @markhoro for putting the new data into context with clear, evidence-based analysis. You highlight that withdrawal is real and often linked to duration of use - reaffirming the need for caution and helping people make informed choices on antidepressants.
The authors found no relationship between duration of prior antidepressant treatment, as stated quite long, and discontinuation symptoms. drugs stopped abruptly. you suggested we missed this, we did not. You like the study, the findings disagree with your premise. Shanti.
In case Dr Jauhar is unsure - there are good estimates of this from the Rosenbaum study (11 months exposure) : 66% for paroxetine and 60% for sertraline. (not a good study to look at fluoxetine) biologicalpsychiatryjournal.com/article/S0006-…
Summarises it better than I could.
🧵Not Everything Is Withdrawal 🚨1/9 When ideology overshadows science, patients suffer. A new systematic review in JAMA Psychiatry just released. Let’s talk about what it shows- and why the current withdrawal ONLY narrative risks harming patients. 1/9👇
I do not think this gentleman has read the paper, as there are a few studies that were longer than eight weeks included in the paper. Shanti.
🧵My response in the New York Times: “If you are looking at people on the drugs for eight weeks, you are not going to find withdrawal,” said James Davies... “It’s like saying cocaine isn’t addictive because we did a study on people who had only been taking it for eight weeks."👇
Antidepressant withdrawal symptoms may be less common than previously thought, a 'gold standard' data review has found. @ImperialBrains, @KingsCollegeLon and @UCL researchers reviewed 50 studies with more than 17,000 participants 🔽 imperial.ac.uk/news/266013/re…
Fairpoint. Though please see this paper, cited in our review. You may know the author. He used this cut off. He just retweeted this, @markhoro
The fact the review is interpreted as ‘reassuring’ by its authors is concerning, and suggests future patients may not be properly informed about the risks of antidepressants 5/n
Theoretically, I would actually agree. However, one cannot place any level of causality on a study sample where response rate was less than 20%, based on a survey.
There is a strong relationship between duration of use and withdrawal effects. We surveyed patients in the NHS and found that people who had used antidepressants for more than 2 years were 10 times more likely to have withdrawal effects…(4/n)
One assume that this newspaper would contact the writers of the original piece, as opposed to critics, who make mistakes perennial?
"Scientists warn antidepressant withdrawal symptoms are being downplayed" - Experts warn that the new review dismisses withdrawal on the basis of 8-week long trials - harming patients in the process. - More critical analysis below👇 independent.co.uk/life-style/hea…
One has to admire persistence. Incorrect persistence, my persistence nonetheless. I assume the journalist does not read the paper either.
"Scientists warn antidepressant withdrawal symptoms are being downplayed" - Experts warn that the new review dismisses withdrawal on the basis of 8-week long trials - harming patients in the process. - More critical analysis below👇 independent.co.uk/life-style/hea…
This makes no pharmacological sense whatsoever. Vnlfxn, paroxetine are the biggest culprits, and affects seen quickly in terms of withdrawal symptoms. Basic.
These conclusions were based a review of short-term trials (8-12 weeks), so of course they would not find much evidence of withdrawal, as withdrawal's incidence, duration and severity is correllated with longer term use (i.e. real world antidepressant use).
Eh? Is this an evidence based approach? No money in antidepressants
It is hard not to think of Naomi Oreskes book ‘Merchants of Doubt’ in which she outlines how academics backed by industry have set out to deny, obscure and minimise health and environmental issues, including links between smoking and cancer. (30/n)
Please read the paper. And @markhoro pointed out a longer-term study by Rosenbaum. He stated we did not include this, we did.
But it only looks at short term use 8- 12 weeks. How is that considered “ balanced”?The study is fundamentally flawed. The vast majority of patients affected by AD withdrawal discontinued drug after a year or more.
Pharmacologically, with paroxetine and vnlfx usually the peak onset of symptoms will be quite quick. Antler data shows v small inc in mean- and large as ie variability. Some people are never happy with findings, which is fine.
It's a limitation of the study. But, in my view, not a major one.