Information about antidepressant withdrawal

I expressed concern about the position statement on antidepressants and depression produced by the Royal College of Psychiatrists (RCPsych) last year (see previous post). The College has now produced an online resource on stopping antidepressants. Even though the document has a disclaimer, I’m not convinced the information for patients is as good as it should be.

For a start, the licensed indications for the various antidepressants on the market do not only include depressive illness, anxiety disorder or obsessive-compulsive disorder. Licences have been granted for other indications that RCPsych does not mention eg. bulimia nervosa.

I think it would have been helpful if the document made clearer that continuation treatment of antidepressants is only really recommended if the medication helps. Some people with depression are not helped within six weeks and there is unlikely to be much point in persisting with the same treatment beyond this period. A change of antidepressant is often tried, and even though no antidepressant has been found to be any better than any other, this change may seem to help. But not everyone benefits from antidepressants, even in the clinical trials.

I can’t find any mention in the document of what I think is probably the most important advice about withdrawal symptoms (which is included in the NICE depression guideline). This is that people should be warned of the risk of discontinuation/withdrawal problems when they first start medication. I commonly used to make a comment like “try not to get too dependent on your medication”. I’ve no evidence that this sort of advice necessarily helps, but I think it stands to reason that it might. People can get psychologically dependent on all sorts of things. Starting antidepressants when one might be feeling desperate and unwell, hoping that the medication may help with such negative feelings, is almost inevitably likely to create a psychological dependence, perhaps particularly if the medication seems to help. When I was working, I always found that patients generally understood what I was saying about the risk of psychological dependence (see previous post).

I’m not quite sure what the evidence is for the position taken in the document that the severity of risk of withdrawal symptoms varies with different antidepressants. For example, as I mention on my ‘Antidepressant discontinuation reactions‘ webpage, fluoxetine was reported to be less likely to cause discontinuation problems, maybe because of its longer half-life, but there was conflict of interest in this work, as it could have had marketing implications for fluoxetine. Perhaps more significantly, no mention is made of the similarity of antidepressant withdrawal with the symptoms of benzodiazepine withdrawal.

Although the document says that the cause of antidepressant discontinuation problems is poorly understand, it does seem to very much favour the idea of re-regulation of receptors. But I’m not sure what the evidence is for this speculation (see eg. previous post). It also uses this speculation to support the argument for tapering, but if this is the explanation for the cause of withdrawal symptoms, then why, for example, is it generally easier to withdraw from antidepressants if they have been prescribed for a shorter period of time? I’m not saying that tapering is not important, which it usually is, but some people do seem to be able to stop suddenly, and the need for tapering also fits with my hypothesis of psychological dependence.

I’m also unclear what the evidence is when the document states that only one third to one half of people experience withdrawal symptoms. For example, in a failed trial of CBT to prevent relapse when withdrawing antidepressants, only 36% of patients succeeded in discontinuing antidepressants over 16 months (see previous post).

Although the document says, “Withdrawal symptoms normally start soon after your medication is reduced or stopped”, I’m not sure if the authors believe in delayed withdrawal problems. Late onset can occur (see previous post).

It is a common misunderstanding that antidepressants take weeks to work (see previous post). Where this comes from is that it generally takes 4-6 weeks for a statistically significant difference between active and placebo treatment to be detected in clinical trials. But this is an artefact of the way in which statistical significance is measured. Larger size clinical trials will detect a statistical difference earlier than trials with smaller numbers of subjects. Actually, the largest improvement per unit time produced by antidepressants occurs within the first 2 weeks of treatment.

There’s no mention in the document about making sure that people close to you are supporting you if you are withdrawing from medication. I think it’s much more difficult to withdraw from antidepressants if your relatives, for example, do not agree that you should be stopping.

I wonder whether it should have been made clearer in the document that gradual tapering does not completely eliminate the risk of withdrawal symptoms. Moreover, taking antidepressants also seems to increase the vulnerability to relapse, as well as cause withdrawal symptoms. That’s why continuation treatment was proposed in the first place.

Personal reflection on critical psychiatry

My invited submission to BJPsych Advances on ‘Clinical reflections on critical psychiatry’ has been rejected. As I’ve mentioned before (see previous post), an advantage of the internet is that these rejected manuscripts don’t have to just be put in a file drawer, never to be seen again. 

The invitation to submit followed a tweet in a conversation with @TheBJPsych suggesting I submit a proposal to BJPsych Advances for a special issue on critical psychiatry. I contacted the journal saying I didn’t know how seriously to consider the suggestion, and the journal answered that it would be pleased to consider a proposal if I would like to submit one, which I duly did. When the editorial board discussed the proposal, it felt that as it already had three special issues in the pipeline for the coming year, it would be better to commission and publish select pieces individually. This led to three of the original contributions to my proposal not being commissioned at all, and at least my contribution to the proposal being rejected.

I have discussed before (see previous post) the difficulties in getting critical psychiatry into mainstream journals. Twitter conversations with @TheBJPsych have tended to be more openminded. It’s even possible that Kam Bhui, as BJPsych editor, had an influence over the acceptance of my recent editorials in BJPsych and BJPsych Bulletin.

Towards a more relational psychiatry

I have mentioned 'relational psychiatry' before (see previous post). Even though I come from a critical psychiatry position, which has never hidden that it grew out of what mainstream psychiatry called 'anti-psychiatry', there are more recent developments from anti-cognitivist phenomenological and enactive accounts of psychopathology that come to the same conclusion about the biomedical model of mental illness. For example, I have mentioned books like Sanneke de Haan's Enactive psychiatry (see eg. previous post) and Thomas Fuch's Ecology of the brain (see another previous post). 

Laurence Kirmayer summarises his perspective on what he calls ‘ecosocial psychiatry’ in an article in World Social Psychiatry. As he says, "Cognitive science supports the view that mental processes are intrinsically social, embodied, and enacted through metaphor, narrative, and discursive practices". As I've argued throughout this blog (eg. see previous post), there needs to be a shift in perspective from a narrow biomedical perspective towards a more truly biopsychosocial approach. As Laurence puts it, what's required is "a shift in perspective from a psychiatry centered on brain circuitry and disorders toward one that recognizes social predicaments as the central focus of clinical concern and social systems or networks as a crucial site for explanation and intervention". In this quote and the article in general, Laurence also helpfully illustrates the importance of social psychiatry, as focusing on the person inevitably means including the interpersonal dimension. This leads, as Laurence says, to an emphasis on "the powerful effects of structural violence and social inequality as key determinants of health".

Reductionism leads to the loss of meaning of human action and a mechanistic psychology cannot be realised in practice (see eg. previous post). This critique of reductionism and positivism in psychiatry, including mechanistic psychological approaches, creates a framework that focuses on the person and has ethical, therapeutic and political implications for clinical practice. It also has consequences for psychiatric research, which has become far too focused on speculative neurobiological notions. 

Descriptive psychopathology is not studied organically at the level of neurobiology. History and mental state examination instead produce a formulation of people’s problems in terms of differential diagnosis and aetiology. Examining the brain in a scanner, for example, does not tell us anything about the cause of thoughts, emotions and behaviour. An integrated understanding of mental dysfunction in the context of the whole person, including emotional needs and life issues, forms the basis for patient-centred and relational psychiatry. The physical disease model of mental illness is outdated and needs to be replaced by a relational psychiatry.

Misdirected aspirations for psychiatric classification

Peter Zachar et al (2019) have published an oral history of the development of DSM-5. For some reason, this does not seem to have included interviewing David Kupfer, the chair of the Task Force, so I do wonder if the article has been affected by the retrospective glosses of Darrel Regier, one of the co-authors of the paper and vice-chair of the Task Force, and Kenneth Kendler, another co-author and chair of the Scientific Review Committee appointed by the President of the American Psychiatric Association (APA). I hadn't realised how much the APA Board of Trustees (BOT) had a role in the DSM-5 process, including appointing Task Force members. 

As Zachar et al note, what was originally envisioned was a shift to a "more scientific basis of psychiatric classification" to take account of the aetiology of mental disorders (see eg. previous post). The paper suggests that any hope for a shift from descriptive to aetiologically-based diagnostic criteria had been largely abandoned by the time the workgroups were finally formed in 2008, not least because the human genome project had not produced candidate genes for mental disorders and there was more questioning about what neuroimaging could achieve.

Darrel Regier reports that DSM-5 leaders were not willing to delay publication, even though the National Institute of Mental Health (NIMH) were about to launch its experimental approach that became the Research Domain Criteria (RDoC) (see previous post). This was because of "what amounted to a resurrection of those very aspirations [that had initially motivated DSM-5]". If this was really the case, then DSM-5 had some foresight, as speculation about neural circuits has gone overboard (see previous post) and Thomas Insel, the previous NIMH director, in my view led NIMH completely 'off beam' with RDoC and so-called precision medicine (see another previous post). 

The paper suggests any decisions about changes were left to the workgroups. To give more direction to the process, BOT created an oversight committee in summer 2009. New guidelines accepted that “DSM-5 will not ‘in itself’ represent a paradigm shift, nor abandon the categorical system of classification, but will start a process that will lead to more useful ways of classifying and diagnosing disorders”. I’m not convinced it really did so, ending up merely with a tinkering with diagnostic criteria far short of the original misguided intentions. It would have been better if the time had been spent rethinking the basis of psychiatric classification (see eg. previous post).

Treatment of depression with antidepressants is primarily a placebo treatment

I’ve mentioned before that NICE may be laying itself open to judicial review about its depression guideline (see previous post). I’ve also emphasised the lack of clear evidence from clinical trials that antidepressants are effective, because placebo amplification may be an explanation of any statistically significant results (see eg. another previous post). This means we should have more of a psychological rather than pharmaceutical model of antidepressant action (Ankarberg & Falkenström, 2008).

Although it should not be surprising, empirical findings confirm that one of the most influential factors in the treatment of depression is the quality of the early therapeutic relationship, not necessarily medication (Blatt & Zuroff, 2005). Pretreatment characteristics of patients may also be factors in outcome. Training in the treatment of depression, therefore, needs to focus on teaching competence in establishing effective therapeutic relationships. Randomised controlled trials, which NICE tends to concentrate on, may be considered the gold standard of experimental design, but naturalistic studies may well have more external validity. The long-term outcome of treatment for depression may not necessarily be that good (see previous post). I’m not encouraging exploitation of the placebo effect, but merely acknowledgement of the importance of the doctor-patient dynamic, even when medication is used (see my BMJ letter). Perhaps NICE should start from a position of therapeutic nihilism before it makes any recommendations about treatment (see last post), but at least it should be clear that the therapeutic relationship is significant and almost certainly affects outcome.

The argument for medical nihilism

Jacob Stegenga published his book on Medical nihilism in 2018. He aligns himself with therapeutic nihilism meaning that "it is impossible to cure people or societies of their ills through treatment” (see Wikipedia entry). This view is contrary to the widespread faith that people tend to have in medical practice. Financial incentives, even corruption, influence medical science. Its research methods are malleable enough to lead to exaggerated claims for effectiveness. We should not be confident about such claims and should be sceptical that medical interventions are effective.

Jacob comes to a position of medical nihilism without apparently fully taking on board the scientîfic challenges in the application of randomised controlled trials (see eg. Kramer & Shapiro, 1984), perhaps particularly the problem of unblinding (see eg. my letter and follow-up). He emphasises the common small effect size of clinical trials, and the fact that some interventions are removed from clinical practice because they are later found to do more harm than good. Few drugs are ‘magic bullets’ in the sense of specifically targeting the cause of a disease. In practice not all clinical trials generally show a benefit for a drug. Bias, even fraud, in clinical research tends to be minimised.

Jacob makes clear he is not saying that no medical intervention is effective. But assessing the effectiveness of medical treatment generally is not merely an empirical matter because of the methodological problems of such research. Research methodology does need to be improved, which will reduce effectiveness estimates, but even so there are still problems about interpreting the data. There is a sense in which it is impossible to be objectively certain about the effectiveness of the vast bulk of treatment. 

Medicine has not really advanced as much as we might like. Broader socioeconomic conditions of health may well be more important than medical treatment itself. In general, there is too much medicine (see eg. post on my personal blog). Jacob encourages medicine to be gentler in its approach, and not so radically aggressive. There needs to be enhanced regulation of medical interventions. The profit motive in medical research is distorting social priorities. The art and science of medicine needs to be rethought (see another post on my personal blog).

“[P]ast 20 years have not been good for the quality of care [in mental health services]”.

BJPsych Bulletin has an interview with Tom Burns, who I have mentioned previously (eg. see previous post). He talks about the OCTET study on community treatment orders (CTOs), which I have also discussed before (see post). Personally I think CTOs should never have been introduced (see eg. another previous post) and Tom now thinks he made a mistake in promoting them. It always strikes me as ironic that there used to be so much concern that a few detained patients were being kept on S17 leave too long, whereas essentially CTOs provide just such a 'long-leash' arrangement. As far as I can see, the introduction of CTOs has led to insufficient use of S17 trial leave. This almost certainly has reduced the opportunity for arranging informal community care without CTO.

Tom is also critical of DSM-5 (see eg. previous post). He's right to focus on descriptive psychopathology as a strength of psychiatry. He calls it 'diagnosis', which can be misleading because history and mental state examination actually lead to a formulation which includes differential (not necessarily a single and certain) diagnosis and aetiology. Diagnosis should not be overemphasised in psychiatric assessment (see eg. previous post). 

Tom's also right about the fragmentation of services over the last 20 years. Hopefully the community mental health framework for adults and older adults will provide a basis for development but it does need leadership to implement it. Despite what Tom says, I haven't abandoned the use of the term 'patient', although also use the term 'service user', even 'survivor', certainly recognising the importance of mental health advocacy (does he?). Tom rightly expresses concern about clustering, which was supposed to support so-called payment by results, which seems to have been quietly dropped (see post on my personal blog). As he also indicates the overpreoccupation with risk has been damaging (see eg. my unpublished article and talk).

A younger version of Tom Burns sounded more optimistic in an e-interview. He even noted the "sense of excitement and 'importance' of psychiatry" created by RD Laing and how anti-psychiatry made the profession "glamorous, albeit controversial". He's still writing about the history of anti-psychiatry (see previous post and my letter).

(With thanks to Suman Fernando who alerted me to the recent interview)