« back · www.benzo.org.uk »
SSRIs, Drug Withdrawal and Abuse:
Problem or Treatment?
Professor C. Heather Ashton
Professor Allan H. Young
University of Newcastle-upon-Tyne, England
First published in:
Selective Serotonin Reuptake Inhibitors (SSRIs):
Past, Present and Future, Chapter 5, 1999
Guidelines for Withdrawal of Antidepressant Drugs
The Ashton Manual · Professor Ashton's Main Page
Selective serotonin reuptake inhibitors (SSRIs) have considerable advantages over earlier antidepressants, such as most tricyclic antidepressants (TCAs) and monoamine oxidase inhibitors (MAOIs), but like all drugs they also have adverse effects. Advantages of SSRIs include: greater tolerability and safety and a wider range of clinical applications, one of which is a potential use in the treatment of drug abuse and some eating disorders (see also: Chapter 4). However, recent evidence shows that SSRIs are associated with a withdrawal reaction on discontinuation after regular use. A further emerging problem is that SSRIs may themselves be entering the repertoire of polydrug abusers. Three aspects of SSRIs are considered in this chapter: withdrawal effects after chronic administration, potential therapeutic value in the treatment of drug abuse and the possibility of SSRI abuse.
THE SSRI WITHDRAWAL (DISCONTINUATION) SYNDROME
Many, if not all, drugs that cause adaptive receptor changes on chronic administration are liable to be associated with symptoms if the drug is abruptly discontinued. Withdrawal symptoms are well documented for TCAs and related compounds, (1-8) MAOIs (6) trazodone (9) and the serotonin (5-hydroxytryptamine, 5-HT) noradrenaline reuptake inhibitor (SNRI), venlafaxine, (10) and it is not surprising that similar symptoms can occur on cessation of long-term treatment with SSRIs. The question of whether the emergence of a withdrawal reaction on drug discontinuation is evidence of drug dependence, as defined for therapeutic dose benzodiazepine dependence, is discussed by Medawar. (11) [See: Social Audit Site owned by Charles Medawar (UK).]
Reactions following SSRI withdrawal have been reviewed by many authors. (10-16) Symptoms appear 1-10 days after stopping or, occasionally, after reducing the dosage of an SSRI that has been taken regularly for a few months or more; the time of emergence depends on the elimination half-life of the individual drug. The symptoms differ qualitatively from the usual side-effects profile of SSRIs and from the illness for which they were prescribed. They are usually mild, lasting for an average of 10 days, although they can occasionally be severe and sometimes persist for a longer period. They respond rapidly to re-administration of the SSRI concerned and may be avoided or minimised by gradual tapering of dosage.
A characteristic cluster of symptoms, or syndrome, is described in various reports (Table 1). Somatic symptoms include: disequilibrium, gastrointestinal symptoms, influenza-like symptoms, sensory disturbances, sleep disturbance and, occasionally, extrapyramidal effects. (17) Psychological symptoms include: anxiety, crying spells, confusion, memory problems, aggression and irritability.
Such symptoms have been recorded after withdrawal of all the SSRIs (paroxetine, sertraline, fluoxetine, fluvoxamine and citalopram). The true incidence is not known, but the relative risk appears to be greatest with paroxetine and least with fluoxetine. (12-15,18-20) Price et al (15) analysed all UK spontaneous adverse drug reaction reports through the 'yellow card' system up to July 1994 and found a 5.1% incidence of withdrawal reactions associated with paroxetine compared with 0.06-0.9% for the other SSRIs. There were 0.3 reports per thousand prescriptions with paroxetine; 0.03 per thousand with sertraline and fluvoxamine and 0.002 with fluoxetine. These proportions agree in general with those reported by Young and Ashton. (18) Both figures are undoubtedly underestimates since they are based on spontaneous reports by doctors, many of whom are unaware of the existence of antidepressant withdrawal symptoms. (21) In small clinical studies involving 6-17 patients, the incidence of withdrawal reactions was 38.5% and 50% for paroxetine (22,23) and 28% for fluvoxamine, (24), while a 34.5% incidence was reported in 55 patients with panic disorder who were withdrawn from paroxetine after 12 weeks of treatment. (25)
Table 1. Symptoms associated with SSRI withdrawal
Somatic Symptoms Psychological Symptoms Disequilibrium:
anorexia, nausea, vomiting,
fatigue, lethargy, myalgia
chills, sweating, comma, headache,
malaise, weakness, palpitations
paraesthesiae, tremor, sensations of
electric shock (often associated with
insomnia, vivid dreams, nightmares
While there is little doubt that withdrawal symptoms can occur when SSRIs are discontinued, their exact mechanisms are far from clear. Several explanations, involving both pharmacodynamic and pharmacokinetic factors, have been suggested to account for the syndrome and for the greater risk associated with paroxetine. (12-14,26) Some symptoms may be at least partially due to cholinergic overactivity resulting from upregulation of muscarinic receptors which occurs in response to chronic use of SSRIs with anticholinergic effects. Withdrawal of the drug reveals the consequent hyperexcitability of cholinergic systems. Such a mechanism has been proposed for the TCA withdrawal syndrome (which has many features in common with that of SSRIs) and is supported by the observation that TCA withdrawal symptoms have been relieved by anticholinergic agents. (2) In this connection it is relevant that, of all the SSRIs, paroxetine has the greatest affinity for muscarinic receptors in the human brain. (27) Withdrawal symptoms due to cholinergic rebound could include gastrointestinal disturbances, influenza-like symptoms, sleep disturbance and mania. An imbalance between cholinergic and dopaminergic activity might account for the occasional appearance of extrapyramidal symptoms. However, Schatzberg et al (26) consider that cholinergic rebound is likely to be a factor only in paroxetine withdrawal symptoms since the other SSRIs have only minimal anticholinergic effects. Even for paroxetine, this explanation may be incomplete: in two reported cases (28) a withdrawal reaction to paroxetine occurred despite treatment with desipramine which binds to the muscarinic cholinergic receptor with approximately equal affinity.
A second factor proposed to account for SSRI withdrawal effects is a decline in serotonergic transmission, although there is no direct evidence for this and the mechanisms for producing particular withdrawal symptoms are obscure. Chronic administration of SSRIs is believed to cause downregulation or desensitisation of inhibitory 5-HT1A-autoreceptors with the result that serotonergic neurotransmission is increased. (29-31) It is hypothesised that following discontinuation of the SSRI this effect is reversed and a relative deficiency of 5-HT in synapses ensues. (26) Decreased serotonergic neurotransmission might account for withdrawal symptoms such as sleep disorder, with rebound of rapid-eye-movement sleep (REMS) and nightmares, (13) impulsive and aggressive behaviour and mood changes. (26) Dizziness, vertigo, nausea and paraesthesiae have also been linked with the role of 5-HT in coordinating sensory and autonomic function with gross motor behaviour. (32) The nature of any such link is obscure but some authors (13,26) point out that such symptoms are often provoked by movement. It is further suggested that extrapyramidal withdrawal symptoms may be related to effects on 5-HT-mediated inhibition of dopaminergic neurotransmission. (10) With regard to individual SSRIs, the most selective SSRI is citalopram, followed by paroxetine which is the most potent. (26) The least selective SSRI is fluoxetine which has some dopamine and noradrenaline reuptake blocking effect. (29) Nevertheless SSRIs may have fewer selective effects in vivo than in vitro tests suggest. Sheline et al (33) found that after 6 weeks treatment with fluvoxamine or fluoxetine in depressed patients, cerebrospinal fluid (CSF) concentrations of the monoamine metabolites 5-hydroxyindoleacetic acid (5-HIAA), 3-methoxy-4- hydroxyphenyl glycol (MHPG) and homovanillic acid (HVA) were reduced by 57%, 48% and 17%, respectively. These results indicate that the drugs, perhaps as a secondary action, affected noradrenergic and dopaminergic neurons as well as serotonergic systems.
A third factor of considerable importance in determining SSRI withdrawal effects concerns their pharmacokinetics. There is a large variation in the rate of elimination between different SSRIs (Table 2). For example, the plasma elimination half-life of paroxetine on chronic dosage is about 21h, while that of fluoxetine is several days. Citalopram and fluvoxamine also have relatively short elimination half-lives, while both sertraline and fluoxetine form pharmacologically active metabolites. The demethylated metabolite of fluoxetine, norfluoxetine, has an elimination half-life of 7-15 days which further prolongs the activity of this compound. Norfluoxetine is further metabolised to a number of other compounds, many of which are unidentified. (14,34)
All the SSRIs are metabolised by the P450 enzyme CYP2D6, which itself shows genetic polymorphism resulting in interindividual variation in rates of metabolism (up to 10% of Caucasians are slow metabolisers of SSRIs) (see also: Chapter 2). Furthermore, paroxetine, and fluoxetine at high plasma concentrations inhibit their own metabolism by CYP2D6 such that they display non-linear pharmacokinetics with a longer elimination half-life at high concentrations, but a more rapid elimination rate as plasma concentrations fall. (12,14) In the case of paroxetine, with its already short half-life, the faster elimination rate, as plasma levels fall following drug discontinuation, may bring about a relatively acute state of cholinergic and serotonergic dysregulation, especially as different receptor adaptations to the drug's presence may reverse at different rates. Possibly this rapid change accounts for the increased prevalence of withdrawal reactions. In contrast, the slow elimination of fluoxetine may allow time for intrinsic readjustment of receptor sensitivities and therefore attenuate withdrawal symptoms.
Table 2. Pharmacokinetic differences between different SSRIs
SSRI Plasma elimination half-life Linearity of
single dose multiple dose
Paroxetine 10h. 21h. Nonlinear Fluvoxamine 11h. 14h. Nonlinear Sertraline 26h. 26h.
Linear Citalopram 33h. 33h. Linear Fluoxetine 1.9 days 5.6 days
Nonlinear: elimination half-life longer at higher plasma concentrations, due to autoinhibition of metabolism; Linear: elimination half-life not dependent on plasma concentration (based on Ref. 14)
Management of Withdrawal
Although the mechanisms of the withdrawal reaction from SSRIs require further explanation, it is clear from clinical evidence that the symptoms can be minimised or avoided by slow tapering of dosage. The principles are similar to those recommended for benzodiazepine withdrawal. (35) Schatzberg et al (16) suggest that paroxetine should be reduced by 5mg/week, and tapering of the shorter half-life SSRIs, especially fluvoxamine and paroxetine, may have to continue for several months. The longer elimination half-life of fluoxetine and norfluoxetine to some extent protects against withdrawal symptoms, but nevertheless reactions can occur even from fluoxetine in some patients. (36) If withdrawal symptoms from any of the SSRIs occur, Lejoyeux et al (10) recommend that the dosage of the drug should be temporarily increased and then tapered again at a slower rate. Although it might appear rational, it is not always possible to substitute one SSRI for another: Lane (1996) (13) quotes the case of a patient who apparently developed withdrawal symptoms three days after switching from paroxetine to sertraline.
Certain patients may be especially vulnerable to withdrawal reactions. These may include patients with anxiety or panic disorders, (25) those who have been on high dosage and those who have had a long duration of treatment. Such patients may require frequent consultations during the course of withdrawal. (10) Ironically, the 10% of slow metabolisers of SSRIs, due to deficiency of the cytochrome P450 CYP2D6 enzyme, may be partially protected from a withdrawal reaction.
On a practical level, slow withdrawal of SSRIs may be difficult due to the limited available dosage strengths in tablet forms of SSRIs. These may not allow a suitable taper and the use of liquid preparations (available for fluoxetine and paroxetine) may be necessary.
SSRIs IN ALCOHOL AND DRUG ABUSE
Pathways of Reward
It is generally accepted that drugs with addictive properties act on brain systems subserving reinforcement or reward. These mechanisms are exceedingly complex and involve both multiple brain areas and multiple neurotransmitters.
One pathway central to reward is the dopaminergic mesocorticolimbic pathway. This originates from dopamine-containing cell bodies in the ventral tegmental area in the midbrain, passes through the medial forebrain bundle and projects to the nucleus accumbens, olfactory tubercle, frontal cortex and septal area. (37) Many addictive drugs activate this system and it has been claimed that it constitutes the final common pathway for all drugs of abuse. (38) Thus cocaine, amphetamine, opioids, nicotine, alcohol, cannabis and other drugs that are misused have all been shown to increase dopamine release in the nucleus accumbens. (39-41) Natural rewarding behaviours, including sexual activity and food reinforcement, are probably also at least partially mediated by this system. (42) Dopaminergic systems probably underlie the positive motivational or incentive aspects of reward and may form the basis of drug-seeking (approach) behaviour. (43)
A second, interacting, reward system utilizing endogenous opioids (ß-endorphin, enkephalins) appears to form the basis for consummatory rewards. (37,43) Although opioids increase dopamine release in the nucleus accumbens, they also subserve reinforcement in animals by a non-dopaminergic mechanism. For example, lesions of the nucleus accumbens and dopamine receptor antagonists drastically reduce cocaine and amphetamine self administration in animals but have much less effect on heroin self-administration. The opioid reward system involves not only the nucleus accumbens but also opioid systems in the periaqueductal grey, amygdala, locus coeruleus and elsewhere. It appears to be largely involved in the consummatory rewards of feeding, drinking, sexual and maternal behaviour. Not only opioid narcotics but also alcohol (44) and possibly benzodiazepines and cannabis have important actions on this system.
A third system postulated to be important for the rewarding actions of sedative/hypnotic drugs is mediated by GABA. (37) Alcohol, barbiturates and benzodiazepines have common actions which include euphoria, disinhibition, anxiety reduction, sedation and hypnosis. In addition, all of these drugs produce a release of punished responding in experimental conflict situations, an effect which correlates well with their clinical actions. This anxiolytic property, mediated by enhancement of GABA activity via interaction with GABAA-benzodiazepine receptors, may be a major component in the rewarding actions and abuse potential of alcohol and other anxiolytic drugs. (45) As well as providing a positive reward, one important factor in their abuse is that they alleviate the anxiety associated with withdrawal from several other drugs of addiction.
Other Neurotransmitter Systems
Many other neurotransmitters are undoubtedly involved in reward systems. These include noradrenaline, which is particularly important in opioid effects on the locus coeruleus, (46) cholecystokinin (CCK; important in signalling satiety), (47) glutamate, neuropeptide Y (41,48) and others, each of which acts on multiple receptor subtypes. The interplay between these complicated systems and those described above remains obscure but may well be different for different drugs and different types of reward.
Interaction of Serotonergic Pathways with Reward Systems
5-HT appears to play a dual role in reward. There is much evidence for an interaction with the mesolimbic dopaminergic pathway (see also: Chapter 10). (49) Both the ventral tegmental area and the nucleus accumbens receive serotonergic projections from the dorsal and median raphe nuclei. Serotonergic activity in the ventral tegmentum appears to be excitatory, resulting in increased dopamine release in the nucleus accumbens. Consistent with this observation, microinfusion of 5-HT into the ventral tegmentum increases responding of rats for a rewarding electrical stimulation of the medial forebrain bundle, suggesting increased activity in the dopaminergic mesocorticolimbic pathway.
Conversely, serotonergic neurons from the raphe nuclei appear to exert an inhibitory effect on dopaminergic neurons in the nucleus accumbens. Thus, lesions of the dorsal and median raphe nuclei in rats increase dopamine turnover in the nucleus accumbens; injection of 5-HT into the nucleus accumbens inhibits the locomotor effects of cocaine and apomorphine, and 5-HT depolarises nucleus accumbens neurons in vitro while dopamine hyperpolarises them.
In summary, serotonergic pathways to the dopaminergic mesolimbic system appear to exert opposing effects, causing excitation in the ventral tegmental area and inhibition in the nucleus accumbens. (49) It is not clear whether the outputs from the dorsal and median raphe nuclei subserve separate functions in the reward-punishment spectrum or whether the opposing effects are mediated by different 5-HT receptors.
There appears to be little information on the interactions between opioids and serotonergic systems (46) but benzodiazepines and alcohol are thought to exert their anxiolytic effects at least partly by decreasing serotonergic activity in critical pathways via GABA enhancement. (50)
5-HT in Addictive Behaviours
In view of these contradictory actions on dopaminergic reward pathways, it is not surprising that the part played by 5-HT in addictive behaviours is uncertain. However, there is some evidence for decreased serotonergic activity in alcoholics, (51) bulimics (52) and possibly in opiate and CNS stimulant abusers, although this may be related to depression. It has been suggested that 5-HT deficiency may underlie drug-seeking behaviour, (51) that it is involved in craving (53) and that brain serotonergic activity contributes to satiety (47,52) and modulates the reinforcing effect or 'high' produced by other drugs of addiction. (54) Trials with drugs that increase serotonergic activity, especially SSRIs, are described below, but it should be noted that these constitute only one of several pharmacological approaches to the treatment of drug addiction. It may be that the combined use of other drugs such as naltrexone, acamprosate, (55) clonidine, lofexidine and others (56) is more effective in some cases.
Several SSRIs including zimelidine, norzimelidine and fluoxetine have been shown to decrease alcohol consumption in alcohol-trained rats in a free-choice environment. (57-59) This effect could occur without significant effects on body weight, total fluid intake or intake of sucrose solution. (58) Drugs with major effects on noradrenaline reuptake (amitriptyline, desipramine, doxepin) did not affect alcohol consumption. (57) Similarly, in rats specially bred for alcohol preference, fluoxetine inhibited intragastric self-administration of alcohol. Treatment with the 5-HT precursor, 5-hydroxytryptophan, (51,60,61) and with 5-HT-releasing agents such as d-fenfluramine (51) likewise reduced alcohol consumption. The magnitude of these effects varied between different studies but, with SSRIs, was generally of the order of 40-50% reduction, both in alcohol preference over water and in total alcohol consumption. (51) In all these investigations the effects of SSRIs were immediate, occurring on the first day of administration.
Conversely, destruction of central serotonergic neurons with the selective neurotoxins, 5,6- or 5,7-dihydroxytryptamine, was reported to enhance alcohol consumption in a free-choice environment, (62) and low doses of the 5-HT1A-receptor agonist, 8-hydroxy2-(di-n-propylamino)-tetralin (8-OH-DPAT), which inhibits 5-HT release through activation of somatodendritic 5-HT1A-autoreceptors, actively enhanced alcohol consumption in rats in a free-choice situation. (63) Furthermore, genetically alcohol-preferring rats, high alcohol-consuming rats and alcohol-preferring mice appear to have reduced central serotonergic function as evidenced by low levels of 5-HT and its metabolite, 5-HIAA, as well as decreased receptor densities in several brain areas, compared with non-alcohol preferring and low-alcohol-consuming rodent lines. (45,64)
Although the above evidence seems fairly consistent in suggesting that low serotonergic activity is associated with increased alcohol consumption and high activity with reduced consumption in rodents, there are some inconsistencies. Most of the drug studies have been limited to short-term drug administration (5-7 days). In one longer-term investigation, Gulley et al (64) studied the time-course of the effects of three SSRIs, fluoxetine, sertraline and paroxetine, on operant lever-pressing for self-administration of alcohol in alcohol-preferring male mice. All the drugs produced initial decreases in lever-pressing for alcohol, but this was followed by a return to baseline over the next few days. After 14 days of treatment, increasing the dose of SSRIs was ineffective in reducing alcohol self-administration. After a washout period of several weeks the drugs again initially decreased alcohol self-administration, followed by a rapid return to baseline. The authors concluded that the effects of SSRIs were related to immediate changes in serotonergic function and that tolerance to this effect developed rapidly.
The role of different types of 5-HT receptors in alcohol consumption is not clear. The effects of SSRIs and 8-OH-DPAT suggest that somatodendritic 5-HT1A-receptors are involved. 5-HT2-receptor antagonists appear to have no effect (51) but 5-HT3-receptor antagonists, such as ondansetron, have been observed to reduce alcohol consumption in alcohol-trained marmosets and in alcohol-preferring rats. (49,51,65,66)
The results from animal studies led to clinical investigations of the effects of drugs which modify serotonergic function in alcoholism. SSRIs, including zimelidine, citalopram, viqualine and fluoxetine have been shown in controlled studies to decrease alcohol consumption in non-depressed alcoholics and heavy or problem drinkers. (67) The effect was dose-related and appeared to require greater than the effective antidepressant dose. For example, fluoxetine at 60mg/day, but not 40mg/day, reduced alcohol consumption in problem drinkers. Body weight and appetite also decreased during SSRI treatment, but the loss of weight was greater than could be accounted for by reduction in calories from alcohol. The consumption of non-alcoholic drinks increased and total fluid consumption was not reduced.
Most human studies with SSRIs have been short-term (2-4 weeks). Unlike the antidepressant action, the effect on alcohol consumption appears to be immediate. In the 28-day study of Naranjo et al (67) there was no difference in the reduction of alcohol consumption compared with baseline between the first and second 14-day periods of treatment with fluoxetine (60mg/day). One open, long-term study of 14 alcohol-dependent patients treated for 6 months with zimelidine (200mg/day) showed a rapid reduction in alcohol intake (within 1 month) with no sign of tolerance over the whole period, but the patients were also receiving psychosocial therapy. (68) In a 3-month controlled study in 108 non-depressed alcoholics, (69) fluvoxamine was found to be superior to placebo in reducing alcohol consumption, with effects apparent at 15 days and remaining significant at 60 and 90 days. Sertraline also appeared to be effective in decreasing alcohol use and improving mood in an open study of 22 depressed alcoholic patients with a history of multiple relapses. (70)
Although these studies showed a statistically significant reduction in alcohol consumption, the effects of SSRIs were modest. In the study of Balldin et al (68) there was no effect on the daily amount of alcohol taken on drinking days, although the number of drinking days per month was reduced from 14 to between 1 and 5 days. In the study of Naranjo et al (67) there was no significant decrease in the number of days of abstinence, but the number of drinks per day was decreased and the total number of drinks per 14-day assessment period was reduced from 115 to 95 drinks, a reduction of 17.3% from the pre-treatment baseline. Overall, studies with SSRIs in alcoholic patients show a reduction in alcohol consumption of only 9-17%. (51)
There is some evidence for reduced serotonergic function in alcoholism. Low concentrations of CSF 5-HIAA and 5-HT, (71-73) decreased whole blood 5-HT concentration, (74) increased platelet 5-HT uptake (75) and increased platelet [3H]imipramine binding (76) have been found in alcoholics whether drinking or abstaining. However, the same abnormalities occur in other conditions, notably depression and impulsive disorders, and in some of the above studies the alcohol-dependent patients also had anxiety and depressed mood. (72-74) Levels of depression were not stated by Daoust et al, (74) while the patients of Patkar (76,77) were described as alcohol-dependent subjects who were "not being treated for depression:" Nevertheless, Sellers et al (51) point out that the effects of SSRIs on alcohol consumption in short-term clinical trials are dose-related and independent of their antidepressant action. Patkar et al (77) also reported a strong positive correlation between craving for alcohol and platelet-rich plasma 5-HT concentrations during detoxification in alcoholic subjects although the relationship between brain serotonergic function and plasma levels of 5-HT was not clear.
Other drugs acting on 5-HT receptors which have been investigated in alcoholic subjects include the 5-HT1A-receptor partial agonist, buspirone, the 5-HT2-receptor antagonist, ritanserin, and the 5-HT3-receptor antagonist, ondansetron. Buspirone was reported to decrease alcohol craving and Hamilton rating scales for anxiety and depression in alcoholic subjects with anxiety disorders. (73,78) Ritanserin reduced craving, anxiety and depressive symptoms during alcohol withdrawal in five alcohol-dependent patients. (79) Ondansetron (0.25mg b.d. and 2mg b.d.) produced a significant reduction in alcohol consumption after 6 weeks of treatment in a placebo-controlled study in alcohol-dependent subjects. (80) The effect was confined to the more moderate drinkers (less than 10 drinks/day) and was more marked with the lower dose. The magnitude of effect was modest, about 18% reduction in average drinking over 6 weeks, similar to the reduction observed with SSRIs.
On the whole, the efficacy of drugs affecting central serotonergic activity as therapeutic agents for alcohol dependence is disappointing. Using, as evaluation criteria, the percentage of continuously abstinent subjects and/or percentage of abstinent days, Zernig et al (81) concluded in a review of the literature over the past 10 years that citalopram, fluoxetine and buspirone were virtually without effect and that acamprosate and naltrexone were the most effective drugs for non-depressed alcohol-dependent patients.
Other Drugs of Abuse
SSRIs have also been shown in some animal and human studies to decrease consumption of other reinforcing drugs including cocaine, amphetamine and opiates. In rats, drugs which modify serotonergic function, including fluoxetine, reduced cocaine and amphetamine self-administration (82,83) and zimelidine decreased morphine consumption in morphine-addicted animals, (84) Antagonists of 5-HT3-receptors did not appear to have similar effects. (85,86) Controlled trials in human drug-abusers are few. In open studies of fluoxetine in cocaine-abusing, heroin addicts entering methadone maintenance programs, cocaine intake and reported craving were reduced in patients taking fluoxetine for at least 1 week. (54,87) The effects appeared to be slow in onset, the steepest decline in consumption occurring at 3 weeks, and to require high dosage of fluoxetine (45-120mg daily); few subjects achieved total abstinence from cocaine. Some patients reported that fluoxetine decreased the quality of the cocaine 'high' and one reported increased rather than decreased craving for cocaine.
Gawin et al (88) carried out a double-blind placebo-controlled study of desipramine and lithium in 72 subjects who abused cocaine only. Desipramine (2.5mg/day) decreased cocaine craving and consumption, 59% of the subjects remaining abstinent for 3-4 consecutive weeks of the 6-week study period, compared with 25% on lithium and 17% on placebo. Similar effects have been reported in open studies with imipramine and trazodone. Batki et al (89) conducted a 12-week placebo-controlled study of fluoxetine (40mg/day) in 32 patients with primary crack-cocaine dependence. The mean dropout rate was significantly greater in the placebo group, only 12% remaining in the study for 6 weeks or more, compared with 68% of those receiving fluoxetine. However, there was no difference in cocaine use or craving between the groups over the first 6 weeks. Other studies cited by Batki et al (89) have shown no benefit from fluoxetine in primary cocaine users or in cocaine users on methadone maintenance.
In amphetamine abusers, Polson et al (90) reported that in a small open study of 13 patients given fluoxetine (20mg/day) for 14 days, five dropped out; four (two of whom were treated longer than 14 days) achieved total abstinence and, in one, there was no change. Seivewright and Carnwath (91) found that fluoxetine (20mg/day) decreased consumption of amphetamine (17 subjects) and cocaine (13 subjects) in primary stimulant users. The effects appeared to be most marked in the 18 patients with depression.
Maremmani et al, (53) noting that opiate antagonists are generally inadequate in preventing relapse in heroin abusers because of continued craving, compared the effects in heroin addicts of a combination of fluoxetine (dose not stated) and naltrexone (9 patients) with those of naltrexone alone (9 patients). In the group taking the combination of drugs only one relapsed over a 3 month period, while five in the naltrexone group relapsed. The authors comment: "Fluoxetine may therefore reduce craving which is the Achilles heel of this condition:"
In general, the evidence indicates that SSRIs may reduce craving in CNS stimulant and opiate abusers, and possibly decrease the drug-related 'high'. However, not all studies have reported benefits and, as with alcohol dependence, the overall effect on consumption and abstinence is modest. It is not clear whether the positive effects are due to an antidepressant action (91,92) or to a more direct effect on the mechanisms underlying addiction, but relatively high dosage and at least several weeks of treatment appear to be necessary. There may be important differences between different SSRIs: Boyer and Feighner (92) point out that fluvoxamine but not fluoxetine increases plasma concentrations of methadone, an effect which may be relevant for polydrug users on methadone maintenance. Prevalence of depression may also be higher in this group than in primary stimulant users.
Abuse of SSRIs
Despite the moderate value of SSRIs in the treatment of drug addiction, there is increasing evidence that these drugs, like other antidepressants, may themselves be abused. In view of the two-edged, stimulant and suppressant, actions of 5-HT on reward systems, this observation is perhaps not as paradoxical as it may at first seem.
It has been long known that addiction to MAOIs, especially those with amphetamine-like structures, can occur with some patients taking large doses to maintain stimulant and euphoric effects. (93) There were also reports of abuse of amitriptyline in opiate users on methadone maintenance programs. Cohen et al (94) reported that 25% of 346 methadone maintenance patients in New York admitted to taking amitriptyline for the purpose of achieving euphoria. Evidence of dependency was deduced from the persistent efforts of many patients to have their dosage increased, attempts to forge prescriptions, the presence of an illicit market for amitriptyline and the confirmation by urinalysis that patients who had not been prescribed it were taking the drug. Cantor (95) confirmed that the practice was not uncommon among opiate-dependent patients and that an active street market for amitriptyline had existed in New York for many years. The effect of amitriptyline taken in doses of 50mg to over 150mg (sometimes up to 20 pills at once) was described as a sedative euphoria and potentiation of methadone effects.
Somewhat later Dorman et al (96) reported misuse of dothiepin among intravenous drug abusers in Dublin: 46% of 83 addicts at a methadone maintenance clinic admitting to misuse of dothiepin in the previous 6 months. Patients described obtaining euphoria and sedation with complex auditory and visual hallucinations which were regarded as pleasant. Dothiepin was taken orally in doses of 150-600mg/day.
The abuse potential of MAOIs and TCAs may not be related to their effects on 5-HT since they also increase synaptic levels of noradrenaline and to some extent dopamine. They may thus have some actions in common with amphetamine which increases central dopaminergic, noradrenergic and serotonergic activity and releases dopamine from the nucleus accumbens. (41) However there is now evidence that SSRIs are also occasionally abused and that they are entering the teenage 'rave scene'. Singh (97) and Singh and Catalan (98) reported the use of fluoxetine and sertraline amongst people taking 3,4 methylenedioxymethamphetamine (MDMA, "Ecstasy") at clubs. Users stated that fluoxetine (20mg) or sertraline (50mg) taken with or before Ecstasy prolonged the 'high' from 2 to 4 hours and made it easier to 'come down' with no hangover. Singh (personal communication) points out that MDMA is largely metabolised by the cytochrome P450, CYP2D6, which is inhibited by fluoxetine and sertraline, and suggests that SSRIs enhance and prolong the effects of MDMA by decreasing its rate of metabolism. However, CYP2D6 inhibition by these SSRIs occurs at high plasma concentrations which take time and regular usage to build up (14) while recreational users take single, and not very high doses, of SSRIs irregularly. The 'high' obtained from Ecstasy is thought to be due to release of 5-HT from neurons arising in the raphe nuclei, (99,100) an effect which in animals and possibly humans (101) leads eventually to 5-HT depletion. Ironically, SSRIs appear to block this effect in laboratory animals (97) and may protect against MDMA-induced neurotoxicity. It is not clear whether SSRIs have similar protective effects in humans since they clearly do not inhibit the Ecstasy 'high'.
Abuse of SSRIs is not confined to Ecstasy users. The Alcohol and Drugs Unit in Newcastle-upon-Tyne (personal communication) confirms the not uncommon use of fluoxetine and paroxetine among young people, usually in combination with amphetamines. Users anecdotally say that these drugs (usually 1-3 tablets) enhance and prolong the amphetamine 'high' and that fluoxetine is better than paroxetine for this purpose. Some users also take amitriptyline, using it mainly as a hypnotic. In this connection, it is interesting to note that fluoxetine has been shown to potentiate the stimulant effects of cocaine in rats, suggesting that it could amplify the subjective effects of cocaine in humans. (102)
These observations suggest that misuse of SSRIs may be a hazard for abusers of Ecstasy, lysergic acid diethylamide (LSD), amphetamine and cocaine. To date there appear to have been no reports of SSRI misuse in opiate abusers. It is difficult to calculate the risks, but Zawertailo et al (103) compared the abuse liability of sertraline, alprazolam and d-amphetamine in 20 volunteers who were experienced but non-dependent users of CNS depressants and concluded that sertraline had a very low abuse potential compared with the other two drugs. Yet it may be salutary to remember that benzodiazepines were once thought to have a low dependence potential but illicit use of these drugs make them (especially oral and intravenous temazepam) the single most abused category of drug in Scotland. (104)
Charney DS, Heninger GR, Sternberg DE et al. Abrupt discontinuation of tricyclic antidepressant drugs: Evidence for noradrenergic hyperactivity. Br J Psychiatry 1982; 141:377-386.
Dilsaver SC, Kronfol Z, Sackellares JC et al. Antidepressant withdrawal syndromes: Evidence supporting the cholinergic overdrive hypothesis. J Clin Psychopharmacol 1983; 3:157-164.
Dilsaver SC, Greden JF, Snider RM. Antidepressant withdrawal syndromes: Phenomenology and physiopathology. Int Clin Psychopharmacol 1987; 2:1-19.
Dilsaver SC, Gredon JF. Antidepressant withdrawal phenomena. Biol Psychiatry 1984; 19:237-256.
Dilsaver SC. Antidepressant withdrawal syndromes: phenomenology and pathophysiology. Acta Psychiatr Scand 1989; 79:113-117.
Tyrer P. Clinical effects of abrupt withdrawal from tricyclic antidepressants and monoamine oxidase inhibitors after long-term treatment. J Affect Disord 1984; 6:1-7.
Drug & Therapeutics Bulletin. Problems when withdrawing antidepressives. Drug Ther Bull 1986; 24:29-30.
Garner EM, Kelly MW, Thompson DF. Tricyclic antidepressant withdrawal syndrome. Ann Pharmacother 1993; 27:1-68.
Otani K, Tanaka O, Kaneko S et al. Mechanisms of the development of trazodone withdrawal symptoms. Int Clin Psychopharmacol 1994; 9:131-133.
Lejoyeux M, Adas J, Mourad I et al. Antidepressant withdrawal syndrome. Recognition, prevention and management. CNS Drugs 1996; 5:278-292.
Medawar C The antidepressant web. Int J Risk & Safety in Med 1997; 10:75-126.
Lane RM. Withdrawal symptoms after discontinuation of selective serotonin reuptake inhibitors (SSRIs). J Serotonin Res 1986; 3:75-83.
Coupland NJ, Bell CJ, Potokar JP. Serotonin reuptake inhibitor withdrawal. J Clin Psychopharmacol. 1996; 16:356-362.
Preskorn SH. Clinical pharmacology of selective serotonin reuptake inhibitors. Professional Communications Ltd., Pfizer, USA.
Price JS, Waller PC, Wood SM et al. A comparison of the post-marketing safety of four selective serotonin re-uptake inhibitors including the investigation of symptoms occurring on withdrawal. Br J Clin Pharmacol 1996; 42:757-763.
Schatzberg AF, Haddad P, Kaplan E et al. Serotonin reuptake inhibitor discontinuation syndrome: A hypothetical definition. J Clin Psychiatry 1997; 58:5-11.
Stoukides JA, Stoukides CA. Extrapyramidal symptoms upon discontinuation of fluoxetine. Am J Psychiatry 1991; 148:1263.
Young AH, Ashton CH. Pharmacokinetics of fluoxetine. Trends Pharmacol Sci 1996;17: 400.
Young AH, Currie A, Ashton CH. Antidepressant withdrawal syndrome. Br J Psychiatry 1997; 170:288.
Committee on Safety of Medicines. Dystonia and withdrawal symptoms with paroxetine (Seroxat). Current Problems in Pharmacovigilance 1993; 19:8.
Young AH, Currie A. Physicians' knowledge of antidepressant withdrawal effects: A survey. j Clin Psychiatry 1997; 58:28-31.
Keuthen NJ, Cyr P, Ricciardi JA et al. Medication withdrawal symptoms in obsessive-compulsive disorder patients treated with paroxetine. J Clin Psychopharmacol 1994; 14:206-207.
Barr LC, Goodman WK, Price LH. Physical symptoms associated with paroxetine discontinuation. Am J Psychiatry 1994; 151:289.
Mallya G, White K, Gunderson C. Is there a serotonergic withdrawal syndrome? Biol Psychiatry 1993; 33:851-852.
Oehrberg S, Christiansen PE, Behnke K et al. Paroxetine in the treatment of panic disorder. Br J Psychiatry 1995; 167: 374-397.
Schatzberg AF, Haddad P, Kaplan E et al. Possible biological mechanisms of the serotonin reuptake inhibitor discontinuation syndrome. J Clin Psychiatry 1997; 58:23-28.
Richelson E. The pharmacology of antidepressants at the synapse: Focus on newer compounds. J Clin Psychiatry 1994; 55 (9 Suppl.): 34-39.
Fava GA, Grandi S. Withdrawal syndrome after paroxetine and sertraline discontinuation. J Clin Psychopharmacol 1995; 15:374-375.
Stanford SC. Prozac: Panacea or puzzle? Trends Pharmacol Sci 1996; 17: 150-154.
Goodwin GM. How do antidepressants affect serotonin receptors? Br J Psychiatry 1996; 164:149-152.
Deakin JFW. A review of clinical efficacy of 5-HT1A agonists in anxiety and depression. J Psychopharmacol 1993; 7:283-289.
Jacobs BL, Fornal CA. 5-HT and motor control: A hypothesis. Trends Neurosci. 1993; 16:346-352.
Sheline YI, Bardgett ME, Csaernansky JG. Correlated reductions in cerebrospinal fluid 5-HIAA and MHPG concentrations after treatment with selective serotonin reuptake inhibitors. J Clin Psychopharmacol 1997; 17:11-14.
Benfield P, Ward A. Fluvoxamine: A review of its pharmacodynamic and therapeutic efficacy in depressive illness. Drugs 1986; 32:313-334.
Ashton H. The treatment of benzodiazepine dependence. Addiction 1994; 89: 1535-1541.
Berlin CS. Fluoxetine withdrawal symptoms. Br J Addict 1996; 84:547-553.
Koob GF. Drugs of abuse: Anatomy, pharmacology and function of reward pathways. Trends Pharmacol Sci 1992; 13:177-84.
Wise RA, Bozarth MA. A psychomotor stimulant theory of addiction. Psychol Rev. 1987; 94:469-492.
Di Chiara G, Imperator A. Drugs abused by humans preferentially increase synaptic dopamine concentrations in the mesolimbic system of freely moving rats. Proc Natl Acad Sci USA 1988; 85:5274-5278.
Miller NS, Gold MS. A hypothesis for a common neurochemical basis for alcohol and drug disorders. Recent Adv in Addictive Disord 1993; 16:105-117.
Nutt DJ. Addiction: Brain mechanisms and their treatment implications. Lancet 1996; 347:31-36.
Philips AG, Fibiger HC. Neuroanatomical bases of intracranial self-stimulation: Untangling the Gordion knot. In: Liebman JM, Cooper SJ, eds. The Neuropharmacologic Basis of Reward. Oxford: Clarendon Press, 1989:66-104.
Di Chiara G, North RA. Neurobiology of opiate abuse. Trends Pharmacol Sci 1992; 13:185-193.
Herz A. Endogenous opioid systems and alcohol addiction. Psychopharmacology 1997; 129:99-111.
Samson HH, Harris RA. Neurobiology of alcohol abuse. Trends Pharmacol Sci 1992; 13:206-211.
Simonato M. The neurochemistry of morphine addiction in the neocortex. Trends Pharmacol Sci 1996; 17:410-415.
Cooper SJ. 5-HT and ingestive behaviour. In: Marsden CA, Heal DJ, eds. Central Serotonin. Receptors and Psychotropic Drugs. London: Blackwell Scientific Publications, 1992: 260-291.
Curzon G, Gibson EL, Oluyomi O. Appetite suppression by commonly used drugs depends on 5-HT receptors but not on 5-HT availability. Trends Pharmacol Sci 1997; 18:21-25.
Tyers MB, Hayes AG. 5-HT receptors and addiction. In: Marsden CA, Heal DJ, eds. Central Serotonin Receptors and Psychotropic Drugs. London: Blackwell Scientific Publications, 1992: 292-305.
Gray JA. Anxiety as a paradigm case of emotion. Br Med Bull 1981; 37:193-197.
Sellers EM, Higgins GA, Sobell MB. 5-HT and alcohol abuse. Trends Pharmacol Sci 1992; 13:69-75.
Kaye WH. Serotonin activity in anorexia and bulimia nervosa. J Psychopharmacol 1992; Meeting of British Association for Psychopharmacology and European Behavioural Pharmacology Society. Abstract 118.
Maremmani I, Castrogiovanni P. Daini L et al. Use of fluoxetine in heroin addiction. Br J Psychiatry. 1992; 160:570-571.
Batki SL, Manfredi LB, Jacob P et al. Fluoxetine for cocaine dependence in methadone maintenance: Quantitative plasma and urine cocaine/benzoylecgonine concentrations. J Clin Psychophamacol 1993; 13: 243-50.
Spanagel R, Zieglgansberger W. Anti-craving compounds for ethanol: New pharmacological tools to study addictive processes. Trends Pharmacol Sci 1997;18:54-59.
Seivewright NA, Greenwood J. What is important in drug misuse treatment? Lancet 1996; 347:373-376.
Amit Z, Sutherland EA, Gill K et al. Zimelidine: A review of its effects on ethanol consumption. Neurosci Biobehav Rev 1984; 8:35-54.
Rockman GE, Amit Z, Brown W et al. An investigation of the mechanisms of action of 5-hydroxytryptamine in the suppression of ethanol intake. Neuropharmacology 1982; 21:341-347.
Murphy JM, Waller MB, Gatto GJ et al. Effects of fluoxetine on the intragastric self-administration of ethanol in the alcohol preferring P line of rats. Alcohol 1988; 5:283-286.
Myers RD, Melchior CL. Alcohol and alcoholism: Role of serotonin. In: Essman BW, ed. Serotonin in Health and Disease, Vol. 2. New York: Spectrum, 1977: 373-340.
Geller I. Effects of para-chlorphenylalanine and 5-hydroxytryptophan on alcohol intake in the rat. Pharmacol Biochem Behav 1973; 1:361-5.
Naranjo CA, Sellers EM, Lawrin MO. Modulation of ethanol intake by serotonin uptake inhibitors. J Clin Psychiatry 1986; 47 (Suppl 4):16-22.
Tomkins DM, Higgins GA, Sellers EM. Low doses of the 5-HT1A agonist 8-hydroxy2-(di-n-propylamino)-tetralin (8-OH DPAT) increase ethanol intake. Psychopharmacology 1994; 115:173-9.
Gulley JM, McNamar C, Barbera TJ et al. Selective serotonin reuptake inhibitors: Effects of chronic treatment on ethanol-reinforced behaviour in mice. Alcohol 1995; 12:177-181.
Silvestre JS, Fernandez AG, Palacios JM et al. Effect of 5-HT receptor agonists and antagonists on ethanol intake and preference in rats. J Psychopharmacol.1995; 9(Suppl): Abstract 211.
Wilson AW, Neill JC, Costall B. Effect of serotonergic compounds on ethanol preference and food consumption in the rat. Psychopharmacology 1995; 9(Supp1): Abstract 212.
Naranjo CA, Kadlec KE, Sanhueza P et al. Fluoxetine differentially alters alcohol intake and other consummatory behaviours in problem drinkers. Clin Pharmacol Ther 1990; 47:490-498.
Balldin J, Berggren U, Bokstrom K et al. Six-month open trial with Zimelidine in alcohol-dependent patients: Reduction in days of alcohol intake. Drug Alcohol Depend 1994; 35:245-8.
Block BA, Holland RL, Ades J. Recidivist alcoholics: A double-blind placebo-controlled study of fluvoxamine. Neuropsychopharmacology 1994; 10(Suppl 3:2):58-43.
O'Brien K. Depression seen during alcohol detoxification: Early intervention with sertraline. Neuropsychopharmacology 1994; 10 No. 3S, Part 2, P-174-5.
.Ballenger JC, Goodwin FK, Major LF et al. Alcohol and central serotonin metabolism in man. Arch Gen Psychiatry 1979; 36:224-227.
Banki CM. Factors in8uencing monoamine metabolites and tryptophan in patients with alcohol dependence. J Neural Transm 1981; 50:89-101.
Tollefson GD. Anxiety and alcoholism: A serotonin link. Br J Psychiatry 1991; 159:34-39.
Banki CM. 5-Hydroxytryptamine content of the whole blood in psychiatric illness and alcoholism. Acta Psychiatr Scand 1978; 57:232-238.
Daoust M, Lhuintre JP, Ernouf D et al. Ethanol intake and 3H-serotonin uptake II: A study in alcoholic patients using platelets 3H-paroxetine binding. Life Sci 1991; 48:1977-1983.
Patkar AA, Naik P, Zaman K et al. Plasma serotonin (5-HT) levels and craving in alcohol dependent subjects during withdrawal. J Psychopharmacol. 1994; Joint Meeting of British Association for Psychopharmacology and Interdisciplinary Society for Biological Psychiatry: Abstract 135.
Patkar AA, Naik P, AI-Chalabi T et al. Platelet imipramine binding in alcohol dependents. J Psychopharmacol. 1994; ]oint Meeting of British Association for Psychopharmacology and Interdisciplinary Society for Biological Psychiatry: Abstract 134.
Bruno F. Buspirone in the treatment of alcoholic patients. Psychopharmacology 1989; 22:49-59.
Monti JM, Alterwain P. Ritanserin decreases alcohol intake in chronic alcoholics. Lancet 1991; 337:360.
Sellers EM, Toneatto T, Romach MK et al. Clinical efficacy of the 5-HT3 antagonist ondansetron in alcohol abuse and dependence. Alcohol Clin Exp Res. 1994; 18:879-885.
Zernig G, Fabisch K, Fabisch H. Pharmacology of alcohol dependence. Trends Pharmacol Sci 1997; 18:229-231.
Richardson N, Roberts D. Fluoxetine pretreatment reduces breaking points on a progressive ratio schedule reinforced by intravenous cocaine self-administration in the rat. Life Sci 1991; 49:833-840.
Carroll M, Lac S, Asencio M et al. Fluoxetine reduces intravenous cocaine self-administration in rats. Pharmacol Biochem Behav 1990; 35:237-244.
Ronnback L, Zeuchner J, Rosengren L et al. Decreased morphine intake by opiate addicted rats administered zimelidine, a 5-HT uptake inhibitor. Psychopharmacology 1984; 82:30-35.
Joharchi N, Sellers EM, Higgins GA. Effect of 5-HT3 receptor antagonists on the discriminative stimulus properties of morphine in rats. Psychopharmacology 1993; 112:111-115.
Hatcher JP, Boyland P, Hagan JJ. The 5-HT3 receptor antagonists, granisetron and ondansetron, do not affect cocaine-induced shifts in intra-cranial self-stimulation thresholds. J Psychopharmacol 1995; 9:342-347.
Pollack MH, Rosenbaum JF. Fluoxetine treatment of cocaine abuse in heroin addicts. J Clin Psychiatry 1989; 52:31-33.
Gawin FH, Kleber HD, Byck R et al. Desipramine facilitation of initial cocaine abstinence. Arch Gen Psychiatry 1989; 46:117-121.
Batki SL, Washburn AM, Frluvvhi K et al. A controlled trial of fluoxetine in crack cocaine dependence. Drug Alcohol Depend 1996; 41:137-142.
Polson RG, Fleming PM, O'Shea JK. Fluoxetine in the treatment of arnphetamine dependence. Human Psychopharmacol 1993; 8:55-58.
Seivewright N, Carnwath T. Fluoxetine in substance misuse treatment. J Psychopharmacol joint Meeting of British Association for Psychopharmacology and Interdisciplinary Society for Biological Psychiatry. 1994; Abstract 139.
Boyer WF, Feighner JP. Other uses of the selective serotonin re-uptake inhibitors. In: Feighner JP, Boyer WF eds. Selective Serotonin Re-uptake Inhibitors (2nd ed). Chichester: John Wiley & Sons, 1996:267-290.
Tyrer PJ. Monoamine oxidase inhibitors and amine precursors. In: Tyrer PJ, ed. Drugs in Psychiatric Practice. London: Butterworth, 1982:249-279.
Cohen MJ, Hanbury R, Stimmel B. Abuse of amitriptyline. JAMA 1978; 240:1372-1373.
Cantor R. Methadone maintenance and amitriptyline. JAMA 1979; 241:2378.
Dorman A, Talbot D, Byrne P et a1. Misuse of dothiepin. Br Med J 1995; 11:1502.
Singh A. Ecstasy and Prozac. New Scientist 1995; Oct:51.
Singh AN, Catalan J. The misuse potential of antidepressants. J Psychopharmacol 1996; 10 (Suppl.): Abstract 122.
Barnes DM. (1988) New data intensify the agony over Ecstasy. Science 1988; 239: 864-856.
Peroutka SJ. Incidence of recreational use of 3,4-methylenedimethoxymethamphetamine (MDMA, 'Ecstasy') on an undergraduate campus. New Engl J Med 1987; 317:1542-1543.
Green AR Goodwin GM. Review of the pharmacology and clinical pharmacology of 3,4-methylenedioxymethamphetamine (MDMA or 'Ecstasy'). Psychopharmacology 1996; 119:247-260.
Cunningham KA, Callahan PM. Monoamine reuptake inhibitors enhance the discriminative state induced by cocaine in the rat. Psychopharmacology 1991; 104:177-178.
Zawertailo LA, Busto U, Kaplan HL et al. Comparative abuse liability of sertraline, alprazolam, and dextroamphetamine in humans. J Clin Psychopharmacol 1995; 15:117-124.
Robertson JR, Ronald PJM. Prescribing benzodiazepines to drug misusers. Lancet 1992; 339:1169-1170.
Professor C. Heather Ashton is Emeritus Professor of Clinical Psychopharmacology and Professor Allan Young is Professor of General Psychiatry at the University of Newcastle-upon-Tyne, England.
Guidelines for Withdrawal of Antidepressant Drugs
C. Heather Ashton, January 2001.
Antidepressant News · Antidepressant Links
The Ashton Manual · Professor Ashton's Main Page
« back · top · www.benzo.org.uk »