Like most other central nervous system drugs, including benzodiazepines, alcohol or heroin, selective serotonin reuptake inhibitor (SSRI) antidepressants may cause withdrawal reactions upon discontinuing the drug, especially after prolonged use (Nielsen et al., Reference Nielsen, Hansen and Gotzsche2012; Chouinard and Chouinard, Reference Chouinard and Chouinard2015). This was first officially acknowledged, but minimised as a minor ‘discontinuation syndrome’, by an industry sponsored consensus panel in the late 1990s, that is, about 10 years after the first SSRI – fluoxetine – was approved as a depression treatment (Schatzberg et al., Reference Schatzberg, Haddad, Kaplan, Lejoyeux, Rosenbaum, Young and Zajecka1997). Since then, withdrawal reactions were sporadically studied and discussed in the scientific literature (e.g. Rosenbaum et al., Reference Rosenbaum, Fava, Hoog, Ascroft and Krebs1998; Haddad, Reference Haddad2001; Baldwin et al., Reference Baldwin, Montgomery, Nil and Lader2007). Common withdrawal symptoms include anxiety, irritability, agitation, dysphoria, insomnia, fatigue, tremor, sweating, shock-like sensations (‘brain zaps’), paraesthesia, vertigo, dizziness, nausea, vomiting, confusion and decreased concentration (Fava et al., Reference Fava, Gatti, Belaise, Guidi and Offidani2015). Although controlled clinical trials and observational studies have revealed remarkably high rates of withdrawal reactions emerging shortly after discontinuation (Rosenbaum et al., Reference Rosenbaum, Fava, Hoog, Ascroft and Krebs1998; Sir et al., Reference Sir, D'souza, Uguz, George, Vahip, Hopwood, Martin, Lam and Burt2005; Fava et al., Reference Fava, Bernardi, Tomba and Rafanelli2007), the preferred narrative in academic psychiatry has always been that withdrawal problems affect only a small minority, are mostly mild and resolve spontaneously within 1–2 weeks (e.g. Burn and Baldwin, Reference Burn and Baldwin2018). This has also been the official position of both the American Psychiatric Association (APA) and National Institute for Health and Care Excellence (NICE) and many leading psychiatrists in the USA and the UK since the early 2000s (National Institute for Health and Care, 2009; American Psychiatric Association, 2010), although, as we will soon state, this position has now changed owing to its being contrary to the evidence base (Davies and Read, Reference Davies and Read2019a; Davies et al., Reference Davies, Read, Hengartner, Cosci, Fava, Chouinard, Van Os, Nardi, Gotzsche, Groot, Offidani, Timimi, Moncrieff, Spada and Guy2019; Horowitz and Taylor, Reference Horowitz and Taylor2019). Below we will therefore outline what conclusions and practical implications can be drawn from the sparse, but vitally important, scientific evidence available.
The first systematic review of SSRI withdrawal reactions was published in 2015 by Fava et al. (Reference Fava, Gatti, Belaise, Guidi and Offidani2015). A few years after their SSRI review, Fava and colleagues also published a systematic review on serotonin-norepinephrine reuptake inhibitor withdrawal (Fava et al., Reference Fava, Benasi, Lucente, Offidani, Cosci and Guidi2018). Two further reviews have recently been published (Jha et al., Reference Jha, Rush and Trivedi2018; Davies and Read, Reference Davies and Read2019a). The four reviews converged on the main finding that the occurrence of withdrawal reactions is quite common, affecting between 30 and 60% of antidepressant users when they try to come off, depending on the methodology deployed (short-term randomised-controlled trials based on pre-selected participants produced somewhat lower estimates than more representative naturalistic studies and large surveys including many long-term users) and on the drugs considered (drugs with a short elimination half-life appear to cause more withdrawal reactions than drugs with a long elimination half-life).
However, there were also points of disagreement between the reviews. For instance, Jha et al. (Reference Jha, Rush and Trivedi2018) claimed that in the vast majority of users, withdrawal symptoms resolve in 2–3 weeks. Their review was not systematic, but selective in the literature it chose to cite – e.g. it was pointed out that their 2–3 weeks statement was not only arbitrary but at odds with the very evidence they cited (Fava and Cosci, Reference Fava and Cosci2019; Hengartner et al., Reference Hengartner, Davies and Read2019). The reviews by Fava et al. (Reference Fava, Gatti, Belaise, Guidi and Offidani2015) and Davies and Read (Reference Davies and Read2019a) were more in accord with each other, showing that the duration of withdrawal symptoms is highly variable, ranging from a few weeks to several months, and occasionally longer. For instance, various studies, using a range of methods, revealed withdrawal durations for over 2 weeks in 55% of patients (Perahia et al., Reference Perahia, Kajdasz, Desaiah and Haddad2005), at least 12 weeks in 25% (RCPsych, 2012), and, in another, for a mean duration of 43 days (Narayan and Haddad, Reference Narayan and Haddad2011). Durations of more than a year are reported in two recent community samples of people experiencing withdrawal reactions – by 38.6% (Davies et al., Reference Davies, Pauli and Montagu2018) and for a mean duration of 90.5 weeks (Stockmann et al., Reference Stockmann, Odegbaro, Timimi and Moncrieff2018). In short, the claim that withdrawal resolves spontaneously in 1–2 or 2–3 weeks as stated in the NICE and APA treatment guidelines, and by Jha et al. (Reference Jha, Rush and Trivedi2018), conflicts with the current evidence base.
Another point of disagreement concerned the incidence of withdrawal reactions when drugs were tapered. While Jha et al. (Reference Jha, Rush and Trivedi2018) claimed that withdrawal occurs mainly when drugs are discontinued abruptly, Fava and colleagues maintained that even the common gradual tapers over a few weeks could not substantially reduce the risk of withdrawal reactions (Fava et al., Reference Fava, Gatti, Belaise, Guidi and Offidani2015; Fava and Cosci, Reference Fava and Cosci2019), a view also supported by recent work published in Lancet Psychiatry, where withdrawal can span over extensive tapering periods (Horowitz and Taylor, Reference Horowitz and Taylor2019). Finally, unlike the other reviews, Davies and Read (Reference Davies and Read2019a) also assessed the severity of withdrawal reactions and found that just under half of all people concerned (46%) rated their withdrawal as severe.
Several reactions to these reviews are noteworthy. However, before we turn to some controversies surrounding antidepressant withdrawal, we first want to point out that there has been a dearth of empirical research on this important issue over the years. As stated above, the first systematic review on withdrawal was not published until 2015. This is remarkable, given that prescribing rates have consistently risen, to alarming levels, over the last 20 years (Davies and Read, Reference Davies and Read2019a), and that almost 200 meta-analyses on the efficacy of new-generation antidepressant have been published between 2007 and 2014 alone, many with industry involvement (Ebrahim et al., Reference Ebrahim, Bance, Athale, Malachowski and Ioannidis2016). Therefore, although it has recently been claimed, by some British psychiatrists, that the psychiatric profession has long recognised the issue of withdrawal (Jauhar et al., Reference Jauhar, Hayes, Goodwin, Baldwin, Cowen and Nutt2019), systematic research into possible harms related to antidepressant discontinuation has obviously been a low priority, for both academic psychiatrists and for the industry.
The public and professional perception of antidepressant withdrawal changed dramatically when Davies and Read (Reference Davies and Read2019a) published their systematic review. In addition to huge media coverage, in particular in the UK, there were also some astonishingly fierce attacks on both the review and on the authors personally by prominent UK psychiatrists (Jauhar and Hayes, Reference Jauhar and Hayes2019; Jauhar et al., Reference Jauhar, Hayes, Goodwin, Baldwin, Cowen and Nutt2019). The attacks contended that the incidence and severity of withdrawal reactions has been exaggerated. They also accused Davies and Read of being biased and partisan. The major critiques of the original review were exposed as misleading or inaccurate (Davies and Read, Reference Davies and Read2019b; Hengartner, Reference Hengartner2019). As for their insinuations about ‘intellectual bias’, we contend that having a clear evidence-based opinion is of lesser concern than the extensive financial conflicts of interests involving drug companies declared in Jauhar et al. (Reference Jauhar, Hayes, Goodwin, Baldwin, Cowen and Nutt2019).
The duration of antidepressant usage has steadily increased over the years (Huijbregts et al., Reference Huijbregts, Hoogendoorn, Slottje, Van Balkom and Batelaan2017; Mars et al., Reference Mars, Heron, Kessler, Davies, Martin, Thomas and Gunnell2017), while large proportions of users indicate that they feel ‘addicted’ to the drugs and experience withdrawal effects (Kessing et al., Reference Kessing, Hansen, Demyttenaere and Bech2005; Read and Williams, Reference Read and Williams2018; Read et al., Reference Read, Cartwright and Gibson2018). Moreover, several clinical trials aimed at discontinuing long-term antidepressant prescriptions failed to successfully withdraw a majority of patients from the drugs despite slow and gradual tapers (Eveleigh et al., Reference Eveleigh, Muskens, Lucassen, Verhaak, Spijker, Van Weel, Oude Voshaar and Speckens2018; Fava and Belaise, Reference Fava and Belaise2018). Yet, academic psychiatry has long clung to the illusion that withdrawal reactions, or discontinuation symptoms, are minor problems that affect only a small minority and which resolve spontaneously within 1–2 weeks, despite a clear lack of supporting evidence (Hengartner and Plöderl, Reference Hengartner and Plöderl2018; Davies and Read, Reference Davies and Read2019a).
We have therefore urged the organisations responsible for depression treatment guidelines to revise their recommendations and to acknowledge that severe withdrawal reactions are much more common than previously believed (Davies et al., Reference Davies, Read, Hengartner, Cosci, Fava, Chouinard, Van Os, Nardi, Gotzsche, Groot, Offidani, Timimi, Moncrieff, Spada and Guy2019). Presumably millions of long-term antidepressant users need help and the difficulties they experience upon discontinuing the drugs need to be detected and correctly diagnosed. Currently, general practitioners who refer to NICE guidelines are likely to misdiagnose withdrawal effects lasting more than 1–2 weeks as the depression returning, and, instead of providing support over a gradual (but individually tailored) withdrawal period, may inappropriately continue or even increase prescriptions.
Fortunately, the tide is finally turning. Firstly, the psychiatric profession has started to acknowledge that these serious issues have long been neglected and minimised. For instance, in May of this year the Royal College of Psychiatrists in the UK published an official statement that severe antidepressant withdrawal needs proper recognition (RCPsych, 2019). Moreover, two psychiatric researchers who personally experienced severe withdrawal, developed a neuropharmacological model of gradual taper to mitigate withdrawal symptoms (Horowitz and Taylor, Reference Horowitz and Taylor2019), and a Dutch team developed tapering strips that help users to withdraw the drugs more safely (Groot and van Os, Reference Groot and Van Os2018). Secondly, and perhaps most significantly, in response to the new evidence we have discussed here, NICE has committed to reviewing its position, held for over 14 years, that antidepressant withdrawal is usually mild resolving over about a week.
We hope that these advances in research and practice will ultimately benefit the millions of antidepressant users who need help. In September of this year researchers, clinicians and ‘experts-by-experience’ from 12 countries will gather in Gothenburg for a meeting of the recently formed International Institute for Psychiatric Drug Withdrawal (http://www.iipdw.com). It does seem that the academic and clinical communities are finally beginning to grapple with the issues that the international online community of antidepressant users (e.g. http://www.letstalkwithdrawal.com; https://www.survivingantidepressants.org) have long been addressing. It is welcome that academic psychiatry, in growing quarters, is finally catching up.
Conflict of interest