Pharmacology of serotonin reuptake inhibitors (SRIs)

Serotonine chemical structure

Not Selective (SRIs is correct, not SSRIs)

These drugs are misnamed; they are not truly selective only for serotonin reuptake. They also affect reuptake of norepinephrine or dopamaine, and have other neurotransmitter effects.  The only agent that is specific only for serotonin among the monoamines is citalopram. The emphasis on serotonin selectivity was a marketing ploy of the 1990s, but it has no scientific basis. Thus the acronym "SSRIs" is scientifically false and should be replaced by the more accurate "SRIs":  these drugs block serotonin reuptake, among other effects.  

Mechanisms of action

All SRIs share a common mechanism of serotonin reuptake blockade, and this effect is essential to the antidepressant effect of this class.  Yet they differ significantly in the amount of serotonin reuptake blockade they produce and they differ in their other biochemical effects.  Thus, I am consciously not using the more common acronym, SSRIs (selective SRIs) to emphasize the point that these agents are not, in fact, particularly selective to the serotonergic system. 

Drug interaction

Perhaps no other class of psychotropic medications has heightened our interest in and knowledge of hepatic cytochrome enzymes.  Differences among SRIs in effects on hepatic metabolic enzymes are also essential to understanding how to use these various medications.   

Sexual dysfunction

Sexual dysfunction is probably the most common side effect of SRIs.  Despite low reports in early controlled clinical trials, later experience indicates that about 50% of patients treated for the long-term with SRIs develop sexual dysfunction.  Dysfunction varies from decreased libido to inability to obtain orgasm and erectile difficulty.  This effect may partly be due to stimulation of 5HT-2 receptors, as agents which block that receptor (like nefazodone and mirtazapine) produce less sexual dysfunction.  Weekend drug holidays, particularly with sertraline as noted below, are sometimes used to minimize this side effect.  It is a clinical reality that sexual dysfunction is underreported if it is not specifically examined by clinicians; patients often will not report sexual side effects due to embarrassment or perhaps due to lack of recognition of those symptoms as medication-related. 

This problem is especially common in SRI treatment of depression since many depressed patients experience sexual dysfunction as part of the depressive syndrome, and it may be difficult to recognize further impairment due to the medications.  A clinician should be suspicious if most depressive symptoms improve except for sexual function.  Again, the clinician needs to consistently query about sexual function to pick up on such lack of improvement.

Other side effects

Other side effects common to the entire class are gastrointestinal side effects and disruption of sleep architecture. 

Gastrointestinal side effects usually involve diarrhea and nausea.  It is an often underrecognized fact that there are more serotonin receptors in the enteric nervous system of the gut than in the brain.  The gut’s enteric nervous system is an independent peripheral nervous system, mostly consisting of serotonergic connections utilizing 5HT-3 receptors.  Mirtazapine, a 5HT-3 blocker, produces fewer gastrointestinal side effects.  Another way to manage such side effects can be to add ondansetron, a selective 5HT-3 receptor antagonist FDA-indicated to treat chemotherapy-related nausea, to the offending SRI.

 The effect of SRIs on sleep architecture (progression from one sleep stage to another) is also common. Sleep stages are highly influenced by serotonergic neurons in the raphe nuclei of the medulla and pons. Progression from one stage of sleep to another is disrupted by SRIs.  Clinically, patients experience mid-cycle awakenings in the middle of the night, as well as vivid dreaming. Trazodone regularizes the transitions between sleep stages, resulting in its popularity as an anti-insomnia adjunct in polypharmacy with SRIs.  Nefazodone and bupropion also appear to improve sleep architecture.

Because of these sleep effects, SRIs should generally be dosed in the morning, with the exception of some persons who experience sedation with paroxetine and occasionally fluoxetine, which should be dosed in the evening in those individuals.

The better than well syndrome
(“Listening to Prozac”)

Prozac (fluoxetine) and SRIs have become famous for possibly making people “better than well”, the concept introduced by Peter Kramer that such SRIs may alter what used to be called “melancholic personality”, a mixture of dysthymic and anxious chronic symptoms.  Such persons seemed to become less anxious and more extroverted with SRIs.    This topic has generated much controversy, both as to whether it really occurs, and if so, what it means ethically and clinically.   After a decade of discussion, my impression is that some patients indeed have such responses to SRIs, which are partly changes in personality and partly improvements in depressive syndromes that had been excluded from DSM-III, such as “neurotic depression”.  On the other hand, these effects of SRIs are not apparently frequent and should always be distinguished from the common occurrence of misdiagnosed bipolar depression with resultant hypomania or mania.

The apathy syndrome
(Listening too much to Prozac)

What is insufficiently appreciated is that the opposite effect can occur as well. Some individuals with a baseline personality that is low in anxiety and highly extroverted (often hyperthymic), will experience the anxiety reducing effects of SRIs as restrictive.     SRIs appear to induce an apathy syndrome in such persons.  This effect is not clearly understood, but it is speculated that SRIs might reduce frontal lobe activity in some patients (although in most SRIs increase frontal lobe activity).  This effect results in a flattening of affect, sometimes described by patients as an decreased ability to “feel” experiences, or an attenuation of normal fluctuations in mood.  In other words, patients may not be able to be appropriately sad when they should be, or appropriately happy when they should be, compared to the manner in which most non-depressed persons would react in those circumstances.  This apathy syndrome can be a subtle effect for patients or clinicians to recognize.  It can be mistaken as a continuation of depressive anhedonia, or possibly seen as a recurrence of depression.  In patients who recover from most of their neurovegetative symptoms except anhedonia, the apathy syndrome should be suspected.  In such persons, I recommend reducing the dose of the SRI or changing treatment to a non-serotonergic agent (like bupropion).

Suicide and akathisia

An unavoidable topic in relation to the SRIs is the risk of suicide.   This issue is most frequently raised in relation to fluoxetine, but this relationship probably reflects the longer availability of fluoxetine relative to other SRIs.  Legal cases arguing for an association with suicide have been raised with most SRIs, and now the FDA has instituted a black box warning of SRI-induced suicidality in children.

The FDA warning is based on a meta-analysis of multiple randomized clinical trials in children, many of which are unpublished, which demonstrated more than 50% increased relative risk of suicidality (attempts or increased suicidal ideation) in a total sample of about 5000 children studied.  While this is important, indicating a real risk, it should also be borne in mind that the absolute rates of suicidality were about 4% with SRIs versus about 2% with placebo, meaning that this increased risk probably occurs in about 5% or less of treated children.  Nonetheless, this fact, as well as the absence of completed suicides in those studies, should not engender false complacency.   As with all randomized clinical trials, the subjects who enter are carefully chosen to be low risk and highly compliant; suicidal ideation of any significance was an exclusion criterion for almost all those studies.   If anything, therefore, the FDA meta-analysis underestimates the true risk in the general population of children.  Further, there are nuisance side effects, and there are fatal side effects.  A 5% risk of nausea is irrelevant, but a 5% risk of death is unacceptable, even if the drug also saves lives (which has not yet been proven with SRIs in relation to preventing suicide). 

Thus, the consequence of the FDA ruling should be what the FDA intended and what has in fact happened:  prior indiscriminate use of SRIs has been replaced by a greater exercise of clinical judgment, which should have always been the case.   The opposite extreme, of never using SRIs, is also inappropriate.

If we now accept that SRI-induced suicidality occurs, the next question is why?   Some critics seem to believe that there is something inherently dangerous about these medications; this claim has not been substantiated.  Two other potential causes seem, in my view, much more likely, and are also preventable.

First, in the opinion of PL, the most likely culprit is misdiagnosed bipolar disorder.  Some studies of children indicate that up to 50% of depressed children (mean age 12) develop manic or hypomanic episodes in a decade of follow-up.  Keeping in mind that the age of onset of bipolar disorder is much earlier than unipolar depression (late teens versus late 20s), clinicians should always have a high index of suspicion for latent bipolar disorder in depressed children.  Further,  manic episodes in children are usually mixed, and about 60% of mixed states involved increased suicidality, a figure that is even higher than the occurrence of suicidality in pure depression.  Statistically, if 50% of apparently depressed children are indeed bipolar, than one would easily expect that 10% of them (or more) would develop manic episodes given antidepressant monotherapy; these figures could easily explain the 5% suicidality rate observed in the FDA database.

Second, SRIs can cause the extrapyramidal symptoms of akathisia.  Akathisia is a very uncomfortable, dysphoric experience, often misinterpreted as agitation or worsening of depression.  When unrecognized and untreated, it can increase suicidal ideation, and seems to have been a factor in a number of the rare cases of fluoxetine-associated suicide.  Though probably less common than SRI-induced mania, SRI-related akathisia has been reported to occur in up to 10% of treated individuals.

What should clinicians do?  Carefully rule out bipolar disorder, keeping in mind that it cannot be effectively ruled out in children since they may not have yet had their first manic or hypomania episode. In children, therefore, the bipolar spectrum concept, including an emphasis on family history, may be especially relevant.  Further, clinicians should warn patients about akathisia and carefully look for this side effect, especially in the first few months of treatment, and be prepared to either reduce the SRI dose, or stop the SRI, or treat with propranolol.  Akathisia should never be left to fester, but needs to be terminated as soon as possible. 

Individual SRIs

Fluoxetine (Prozac)

The groundbreaking medication in this class was fluoxetine, introduced in the United States in 1989.  This agent was quickly followed by sertraline, paroxetine, fluvoxamine, and citalopram, in that order.  Fluoxetine is also the first of these agents to become available in generic form (2001), after more than a decade of blockbusting profits.  In my opinion, it is not the best SRI, but it had the major advantage of being the first on the market. Both clinicians and patients quickly became familiar, and usually comfortable, with this medication.  By initiating the change in prescriptions from TCAs to newer antidepressants, fluoxetine also became the symbol of the new generation of kindler, gentler psychiatric medications.  Patients and clinicians began to “listen” to it for all kinds of benefits and it is still not unusual to see patients come to a clinician’s office specifically asking for this agent.

From a sober medical perspective a decade later, however, fluoxetine has some advantages and some disadvantages that do not seem to distinguish it from its SRI compatriots. 

Perhaps its main advantage, along with citalopram and sertraline, is that it has become available in the US in a generic formulation; cost should no longer be a limiting factor in the use of these agents.

A unique feature of fluoxetine is that it and its active metabolite norfluoxetine have quite long half-lives, the longest, in fact, of any major psychotropic agent.  The half-life of fluoxetine is about one day and of norfluoxetine is about 3-5 days.  Thus, on average, it takes 4 days for a dose of fluoxetine to be 50% eliminated.  Since it takes 3 half-lives to achieve a steady-state blood level, it takes 12 days just to achieve a steady-state level of this agent.  All antidepressants require about 4-8 weeks time delay for their pharmacological effects to translate into a clinical antidepressant effect.  This delay probably reflects intracellular second messenger and genetic changes.  However, this 4-6 week delay occurs after steady state is achieved.  For most antidepressants, a steady state blood level is achieved in 1-2 days.  For fluoxetine, the 12-day delay on average means that the clinical effect of the medication may be delayed another 1-2 weeks.  This is why fluoxetine is the only antidepressant that requires a 6-8 week period for a full therapeutic trial (as opposed to 4-6 weeks).  This fact may be a partial disadvantage: if a patient is not responding to fluoxetine at 4 weeks of treatment, one cannot change to another medication knowing that a full trial has been given unless one waits another two weeks.  This is not the case with other antidepressants, in which no response at all at 4 weeks is a sufficient therapeutic trial.  On the flip side, a potential advantage to this long half-life is that fluoxetine does not leave the body quickly, and thus may be less prone to causing serotonin withdrawal syndrome.

Another feature of fluoxetine is that, contrary to what is sometimes assumed, it is not purely selective for serotonin.  In fact fluoxetine mildly blocks the reuptake of norepinephrine.  This effect of fluoxetine is not minor, and is somewhat similar to the effect of venlafaxine.  This feature may possibly account for the “stimulating” effects frequently reported with fluoxetine.

Like all SRIs, fluoxetine can disrupt the sleep architecture.  Combined with its potentially stimulating effects, fluoxetine can cause insomnia.  (Nonetheless, a small subgroup of patients actually become sedated with it).

Another major effect of fluoxetine is that it is a strong inhibitor of most hepatic cytochrome P450 enzymes, thereby increasing the blood levels and effects of many other medications, including neuroleptics, TCAs, and some mood stabilizers.

Despite its side effect of sometimes causing “stimulation,” fluoxetine often has the pharmacological effect of decreased anxiety, as do all SRIs. 

Since fluoxetine has been available for the longest amount of time, it has been studied in the largest number of conditions, with reports of efficacy in bulimia nervosa, anorexia nervosa, PTSD, personality disorders, OCD, and panic disorder.  It has efficacy in these conditions, but all SRIs are likely useful in similar conditions. Fluoxetine is now FDA-indicated (under a different trade name) for the treatment of premenstrual syndrome (late luteal phase dysphoric disorder).

As noted above, some clinicians believe that fluoxetine exerts a special beneficial effect on personality.  This perspective holds that some depressed and even non-depressed individuals become “better than well” on fluoxetine; that is, not only are they no longer depressed, but they also do not return to their premorbid personality.  They often become more extroverted and fun-loving, and a patient may feel that her personality on fluoxetine is her “real” self.  When reported, this effect seems to be rare, and it seems not to be clearly specific to fluoxetine.  It might be explained as an effect on personality, separate from any antidepressant effects of the medication.  Some researchers have provided evidence that some aspects of personality, especially what has been called “harm avoidance”, may be associated with serotonergic areas of the brain.  By increasing serotonin availability, SRIs may alter personality by making persons less harm avoidant, that is, less cautious or shy or introverted.  As a result, this at one level mysterious effect of fluoxetine may simply be a straightforward effect on the biochemistry of personality.  At another level, however, this effect of fluoxetine needs to be balanced with an apparently equal potential for less positive effects on personality, such as the apathy syndrome.   

Sertraline (Zoloft)

Sertraline, now also available as a generic agent in the US, is an SRI with a moderate amount of dopamine reuptake blockade.  It has a shorter half-life than fluoxetine, about one day, and causes much less inhibition of hepatic cytochrome P450 enzymes.  This does not mean it has no hepatic enzyme effect; especially at higher doses, its effect can be clinically notable.  Usually, though, its hepatic effects are mild and do not lead to clinically significant drug interactions.

Sertraline shares with other SRIs a general antianxiety effect, as well as a potentially disruptive sleep effect.  It is now FDA indicated for the treatment of PTSD.  While fluoxetine possesses an FDA indication for premenstrual syndrome, sertraline is also useful in this setting, and perhaps more so if an individual prefers to use the SRI for 5 days before and after the menstrual period, rather than continuously.  Such short term use is effective with all SRIs except fluoxetine (again due to the latter’s long half-life).  The short half-life of sertraline may also be useful in the context of drug holidays for sexual dysfunction. 

This discontinuation would not work with fluoxetine due to its long half-life, and it might be more difficult with paroxetine or venlafaxine, due to their short half-lives and the consequent risk of serotonin withdrawal syndrome.  The weekend drug holiday seems to work well with sertraline, however, which may partly reflect the fact that sertraline’s metabolite, desmethylsertraline, has a half-life of about 3 days.  As a result, in weekend drug holidays, sertraline is short-acting enough to leave the body such that sexual function temporarily improves, but not so short-acting that serotonin withdrawal symptoms begin.

A potential disadvantage of sertraline’s mild dopaminergic effect may be that in particular susceptible individuals, this biochemical property may lead to increased psychosis.  This has been reported in otherwise nonpsychotic individuals, though infrequently.  In bipolar disorder, sertraline has not seemed clinically useful in my experience, with quite notable mania induction, though empirical data are not available.  Sometimes I am asked why I think sertraline’s dopaminergic effect is a problem in terms of increased mania risk, while bupropion possesses less mania risk.  It may be a matter of degree, since sertraline, contrary to common belief, has more dopaminergic effect than bupropion.  These mechanistic issues are really a matter of speculation.  The fact remains that at a clinical level, in some persons, sertraline appears prone to inducing psychosis or mania.

Paroxetine (Paxil)

Paroxetine is also short-acting, with a half-life of about one day, and perhaps more short-acting than sertraline (since the latter has a metabolite with a long half-life).  Paroxetine is more potent than either sertraline or fluoxetine in its serotonin reuptake effect, which means that if a higher degree of serotonin reuptake is needed, then paroxetine may be useful.  In other words, even strictly on the issue of serotonin reuptake effect, these SRIs are not identical.  This difference may account for the clinical experience that some patients respond to one SRI and others to another SRI. 

Paroxetine also possess moderate anticholinergic effects.  While prominent in vitro, in vivo studies in humans suggest that these anticholinergic effects are not marked, and are certainly less severe than with TCAs.  In susceptible persons, though, clinically notable anticholinergic effects may occur. These include dry mouth, sedation, constipation, and cognitive side effects. 

Most SRIs are weight neutral, with some persons losing weight and a small number of persons gaining weight.  Of the SRIs, paroxetine appears to be the most liable to some amount of weight gain, though, not, on the whole, to a severe degree.  Some individuals will experience significant weight gain.

Among the SRIs, which all have anti-anxiety effects, paroxetine holds the reputation of having the most consistent anxiolytic effects, and it is now FDA-indicated in the treatment of generalized anxiety disorder, panic disorder, and social phobia (social anxiety disorder). 

Unfortunately, paroxetine’s short half-life makes it somewhat more liable to serotonin withdrawal symptoms than other SRIs.  Again, these symptoms are finite and usually tolerable with appropriate clinician support.

Paroxetine has minimal effect on the cytochrome P450 2D6 system, unlike fluoxetine, but is a strong inhibitor of the cytochrome P450 3A4 system, like fluoxetine.  In terms of drug interactions, paroxetine has an intermediate effect—not as marked as fluoxetine, but not as mild as sertraline.

Citalopram (Celexa)

Citalopram, now available in generic form in the US, was used in Europe for a number of years before it was introduced into the United States recently in 1999.  (In fact, in Europe it predated fluoxetine). It is even more potent in its serotonin reuptake blockade effect than paroxetine.  It also the most purely serotonergic agent in this class, with almost no other effects on other neurotransmitter systems.  It has minimal effect on hepatic enzymes, and a short (but not too short) half-life of about 1 day.  In many ways, then, citalopram may deserve the label of the classic SRI.

Overall, it likely has similar benefits to most SRIs as regards antianxiety and other effects.  Due to its “cleaner” biochemical profile, it may particularly helpful in elderly patients, in whom it is less likely to cause avoidable side effects or drug interactions. 

It should not be given above 40 mg/d due to risk of cardiac arrhythmia.

Escitalopram (Lexapro) is the active enantiomer of citalopram; except for proving more profits to its makers, and the ability to get the same effect of citalopram at lower doses, I see little need to use this expensive agent.  Its maker’s claims to better tolerability than citalopram have not been clinically confirmed.

Fluvoxamine (Luvox)

Fluvoxamine is FDA-indicated for obsessive-compulsive disorder, but likely has benefits for depression and anxiety similar to other SRIs.  Like paroxetine and citalopram, it is potent in its serotonin reuptake blockade.  It has few other biochemical effects, and no other real  advantage to other SRIs.  It is a strong inhibitor of the cytochrome P450 3A4 system, perhaps even more than paroxetine, and thus has some disadvantages in terms of drug interactions. 

Typical doses and features of specific agents

  • Fluoxetine (Prozac) 20-80 mg/d  Longest half-life of any antidepressant, requiring longer duration for a therapeutic trial but also producing less serotonin withdrawal, somewhat noradrenergic, marked drug interactions (all cytochromes, esp 2D6 & 3A4) 
  • Sertraline (Zoloft) 50-200 Dopaminergic, useful in weekend holidays for sexual dysfunction, mild drug interactions 
  • Paroxetine (Paxil) 20-50 Very anxiolytic, moderately anticholinergic, some weight gain, somewhat more serotonin withdrawal, multiple FDA indications, cytochrome 3A4 inhibitor
  • Citalopram (Celexa) 10-40 Most potent serotonin reuptake inhibition, most selective agent for serotonin, minimal drug interactions, especially useful in elderly
  • Fluvoxamine (Luvox) 50-250 Highly potent serotonin reuptake inhibition, indicated for OCD

General rules for SRI treatment

  1. Dose all agents once daily in the morning, with the exception of paroxetine (which can be sedating) and fluoxetine (in the minority of persons who experience sedation)
  2. Fluoxetine requires a longer therapeutic trial, but has less serotonin withdrawal
  3. Sertraline works well in weekend drug holidays for sexual dysfunction 
  4. Citalopram is the most selective for serotonin reuptake 
  5. Sertraline and citalopram have the fewest drug interactions. 
  6. Paroxetine may have the most anxiolytic effect 
  7. Observe carefully for mixed states in misdiagnosed bipolar disorder, as well as akathisia, as potential risk factors for suicidality

Meet our expert EDITORIAL BOARD, composed of clinicians and researchers from around the world. 

Subscribe to the RSS feed below to follow our "What's new" blog posts