Antidepressant efficacy and side-effect burden: a quick guide for clinicians

TCAs and MAOIs

These two classes of compounds comprise the earliest specific ADT. Their clinical introduction in the 1950s following fortuitous discovery of their antidepressant properties marked the advent of psychopharmacotherapy as an indispensable tool in the treatment of MDD and spurred the first monoaminergic theories in the etiopathogenesis of depression.

TCAs, which take their name from the basic three-ring chemical structure common to them, act primarily by elevating serotonin and norepinephrine levels via uptake inhibition (similar to the later developed serotonin norepinephrine reuptake inhibitors [SNRIs] mechanistically). However, as they also antagonize muscarinic acetylcholine receptors, they are prone to anticholinergic side effects (e.g., dry mouth, blurry vision, constipation, urinary retention), which often limit their utility. In addition, TCAs are known to cause prominent weight gain and sedation and can block cardiac sodium channels, which in the case of overdose may lead to sudden cardiac death [19]. In a meta-analytic review of TCA compared with SSRI side effects, Montgomery et al. found that patients on TCAs discontinued treatment 27% of the time compared with 19% for those on SSRIs [20]. In elderly MDD patients, the rates were 33% and 16%, respectively [21].

MAOIs, in contrast, act by inhibiting the activity of the enzyme monoamine oxidase, thereby preventing the breakdown of monoamine neurotransmitters. Two enzyme isoforms exist, MAO-A and MAO-B, which preferentially degrade different amines. The early nonselective MAOIs, like the TCAs, were often limited in their use due to adverse events. In the case of MAOIs, this included dangerous and potentially lethal interactions with food, particularly foods rich in tyramine (e.g., aged cheese), and with other medications. Fatal serotonin syndromes or hypertensive crises may develop, respectively, by inappropriate use of these agents. In fact, MAOIs should not be used in ADT augmentation with SSRIs due to a potentially lethal increase in serotonin, known as “serotonin syndrome.” MAOIs are also known to promote weight gain, as well as fatigue and hypotension. Consequently, they are often the last pharmacologic alternative after all other ADT options have failed to yield remission. Some newer MAOIs, however, such as selegiline and the reversible MAOI moclobemide, have proven safer and may be considered for use earlier in treatment perhaps.

Thase et al. conducted a meta-analysis of all published reports comparing TCA and MAOI agents [22]. This team found that, although MAOIs outperformed placebo in treating hospitalized, severely depressed patients, they were significantly less effective than TCAs in this patient subgroup. In contrast, other studies have found that MAOIs are more effective than TCAs in treating outpatients who were less depressed and exhibited more atypical MDD features [23]. Such findings suggest that these two classes of ADTs may have a niche, or specific patient subtype, for which they are most effective, a fact which further underlines the importance of considering inpatient–outpatient status and depressive severity in treatment algorithms. RCT data equating these two ADT classes in the absence of stratified analysis should therefore be heavily scrutinized by the prescriber. Meta-analyses comparing these classic ADTs with the more novel (post-1980s) ADTs have also been published. According to these data, the advantage of TCAs in the more melancholic and severely depressed inpatient setting persists not only over MAOIs but also over the well-prescribed, modern SSRIs [18,22,24–27]. The TCA agents imipramine and amitriptyline faired the best. Considering the mechanism of action of the TCAs, it is postulated that serotonergic facilitation ameliorates limbic dysfunction, whereas more norepinephrine promotes better function in the dorsolateral prefrontal cortex (DLPFC).

While these two older classes have largely fallen out of favor, there continues to be wide-scale use of TCAs for insomnia, pain, and resistant MDD, and of MAOIs for resistant MDD and social anxiety disorder. With regard to the MAOIs, there is some new research to indicate that the risk and severity of adverse dietary interactions may be much less alarming than previously thought, especially in the setting of proper monitoring and adherence [28]. Nevertheless, misconceptions and fears of the drug’s risks can lead to patients never starting or later discontinuing potentially helpful treatment. This tends to promote a conservative prescribing practice. It is, therefore, important that clinicians address and discuss with their patients any fears or concerns about medication side effects in order to keep all therapeutic avenues open.

SSRIs

The SSRI class of ADT is the most widely utilized in the modern era of MDD treatment. It is comprised of five agents: fluoxetine, sertraline, paroxetine, citalopram, and escitalopram. These agents use the serotonin reuptake inhibition property found in a majority of the TCA class, but are selective for this one mechanism alone, and thus avoid many of the anticholinergic and cardiac side effects of the TCAs. Additionally, they do not require dietary and drug-related restrictions as the MAOIs do. The enhanced safety of SSRIs drove a veritable revolution in the treatment of depression during the 1990s, allowing more patients to be treated by prescription than ever before. During the previous era, MDD patients had to suffer severe MDD or melancholia to warrant the side-effect risk associated with these ADTs. With the development and introduction of safer drugs on the market, and the risk of pharmacotherapy greatly reduced, ADT could now be prescribed for MDD patients who were mild or moderately symptomatic, possibly even subsyndromal [29].

Having now introduced the most commonly prescribed first-line treatment for MDD, it will be useful here to provide the reader with a basic neurobiological picture of how a typical SSRI may start to relieve depressive symptoms in a patient with moderate, uncomplicated MDD. As mentioned, a typical SSRI blocks serotonin reuptake pumps (more often called serotonin transporters in the literature), acutely raising this transmitter in neuronal synapses. Despite this relatively immediate effect, MDD symptoms usually begin to resolve a few weeks later [30], suggesting that an acute rise in the level of neurotransmitter is necessary but insufficient to effect a change in phenotypic, exhibited depressive symptoms. For observable symptom improvement to arise, there must be a gradual shift in brain homeostasis and neurofunctioning over time, a shift away from a current depressive baseline, involving many different neuroanatomical areas and structures in the brain (e.g., receptors and enzymes). In this process, elevated serotonin may be, at first, biologically “interpreted” by neurons as locally toxic, causing certain cells to undergo stress-related changes in order to accommodate, or adapt to, higher-than-baseline serotonin levels. Subsequently, the activity of certain neuronal enzymes (i.e., those involved in serotonin metabolism) will increase, while that of others will decrease; similarly, a subset of receptors will start to become desensitized or downregulated, while other receptors (i.e., those responsive to serotonin) will be more heavily synthesized and shuttled to the plasma membrane. This degree of selective gene activation with associated protein formation may take 4–6 weeks, thus correlating with the time it takes an antidepressant to reach full effect [30].

According to this view, then, antidepressants create a new environment in the central nervous system (CNS)—a favorable yet initially stressful chemical milieu—that prompts certain genes to be activated, and new proteins to be synthesized, that may better accommodate the perceived neurochemical stress. However, once these changes have taken place, sustained increases in serotonin associated with SSRI adherence will start to reestablish a new, elevated baseline, one which will not only allow a return to normal functioning but also provide the neurotrophic substrate necessary for building new adaptive neural pathways. In fact, many antidepressants and mood stabilizers, regardless of mechanism of action, promote greater amounts and activity of brain neurotrophic factors, which may reverse MDD-induced brain atrophy, increase synaptic plasticity, and perhaps restore balance between limbic and cortical activities to alleviate depression [31–33]. This ability to increase gene activity to promote better synaptic levels of neurotrophic proteins may indeed start with the excessive “toxic” serotonin initially provided, most commonly, by SSRI monotherapy.

Among the SSRI class, there are some subtle yet notable pharmacologic differences. For example, the therapeutic effects of fluoxetine may emerge more slowly than those of some other SSRIs. Escitalopram may be more effective than its parent drug citalopram [34]. While all of the mentioned SSRIs are approved to treat MDD, some are also specifically approved to treat panic disorder, social anxiety disorder, obsessive– compulsive disorder, post-traumatic stress disorder (PTSD), bulimia nervosa, and/or premenstrual dysphoric disorder. To be FDA approved for a particular disorder, an SSRI must have at least two positive RCTs demonstrating their efficacy against that disease. Sertraline and paroxetine have the most FDA approvals and indications to treat a myriad of psychiatric disorders. An initial SSRI monotherapy should be chosen based on the patient’s diagnosis utilizing knowledge of these FDA approvals [23].

There are also subtle side-effect differences among the SSRIs. All may cause headaches, gastrointestinal disturbance, insomnia, fatigue, initial anxiety, and so on. Paroxetine eventually may allow more weight gain. This and citalopram may be the most sedating. Sertraline may have more adverse gastrointestinal effects. This and paroxetine more commonly cause withdrawal symptoms upon abrupt cessation due to their relatively short half-lives. Paroxetine and fluoxetine have more drug–drug interactions via CYP 4502D6 hepatic enzyme inhibition. These subtle diagnostic considerations, differences in efficacy, and adverse effects are usually weighed by the prescriber prior to choosing the ADT on a per-patient basis. Each SSRI is felt to be unique in clinical application despite being considered essentially equal per current regulatory processes.

SNRIs

Other modern ADTs exist that are not SSRI based. These include the SNRIs—chiefly venlafaxine, desvenlafaxine, duloxetine, and levomilnacipran—which dually inhibit serotonin and norepinephrine reuptake pumps comparably, allowing treatment of a wide range of depressive symptoms. This kind of dual reuptake inhibition is similar to the TCAs, but like the SSRI class, SNRIs are associated with less serious side effects. Some of these side effects include initial increases in anxiety, insomnia, and restlessness, and possible sexual dysfunction and headaches as well. Compared with the SSRI class, the SNRI class tends to induce more nausea, insomnia, dry mouth, and in rare cases elevated blood pressure [23].

With regard to comparative efficacy, Thase and colleagues [35] conducted a pooled analysis using the more stringent depressive remission benchmark over the usual FDA 50% response benchmark. They found that 45% of venlafaxine-treated patients gained remission compared with 35% on SSRI and 25% on placebo [35]. This was largely replicated by Nemeroff et al. in his team’s meta-analysis and also by a smaller study by Smith et al., all observing superior response rates for SNRIs [36,37]. A pooled analysis of another SNRI, duloxetine, utilizing less RCT data, found no statistical differences with SSRI [38] despite better response rates and higher HAM-D scale scores. It is currently unclear, therefore, whether SNRIs globally achieve a greater effect in MDD treatment or whether the effect is drug or dose dependent.

Desvenlafaxine, synthesized from the primary active metabolite of venlafaxine, is a relatively new SNRI developed to improve upon the strength of its parent drug. In addition to in vitro studies indicating its more potent inhibition of norepinephrine transporters, noninferiority RCTs have compared the efficacy (symptom reduction) of desvenlafaxine favorably with venlafaxine and SSRIs. This is evidenced by a similar magnitude of change among experimental groups from baseline to endpoint on mean HAM-D₁₇ scores compared with placebo [39,40]. Treatment outcome (effectiveness) data have also been positive, with observed remission rates similar to those reported by Thase and colleagues from placebo-controlled venlafaxine studies [41].

The fourth member of the SNRI class, levomilnacipran, differs from the other SNRIs in being a more potent and selective inhibitor of norepinephrine than serotonin, especially at low doses [42]. In the absence of modern selective norepinephrine reuptake inhibitors, this characteristic may represent a novel therapeutic contribution to MDD patients, especially among those with symptoms of DLPFC hypofunction (e.g., fatigue, poor concentration, executive dysfunction) who have not been able to tolerate other SNRIs.

Mirtazapine, trazodone, and bupropion

Among the ADTs arsenal prescribers have at their disposal, some drugs have been labeled “atypical” in the sense that they do not fit neatly—whether structurally or mechanistically—into any of the previous classes. Mirtazapine is one such agent, named colloquially one of the “sedating antidepressants,” as it tends to promote sleep or drowsiness. Unfortunately, its daytime sedation rates are high, and it is a weight gain-prone ADT. It has a low rate of sexual dysfunction [23]. It does not utilize a typical SNRI mechanism, but rather, in a more complicated manner, inhibits norepinephrine alpha-2 autoreceptors, allowing more norepinephrine to be released from nerve terminals. It also blocks 5-HT2A/2C receptors, thus allowing more serotonin, dopamine, and norepinephrine modulation in the cortex. It may be described in more specific terms as a noradrenergic antagonist-specific serotonin antagonist (NaSSA). As such, it achieves greater neurotransmitter levels via a different mechanism of action than the SNRI or SSRI, perhaps also explaining its different side-effect profile. Quitkin’s pooled analysis showed a faster onset of effect with mirtazapine yet no ultimate significant difference in effectiveness between mirtazapine and the SSRI [43]. Further meta-analysis has confirmed the accelerant property of mirtazapine monotherapy, suggesting faster onset of efficacy but equal efficacy by the end of 8-week acute trials [44].

Like the SSRIs, trazodone was an early “second generation” ADT whose low side effect and toxicity profile made it a popular alternative to MAOIs and TCAs in the early 1980s [45]. As an atypical agent, its mechanism of action is complex and multifaceted. Marketed as a mild serotonin reuptake inhibitor, trazodone, in fact, can be viewed as a mixed serotonergic agonist–antagonist and is more widely referred to by clinicians as a serotonin antagonist and reuptake inhibitor (SARI). The drug increases serotonin levels in the CNS through a combined effect on serotonin reuptake pumps and 5-HT2A/2C receptors via both receptor agonist and antagonist activities [46,47]. It is considered, like mirtazapine, a sedating ADT, and marked sedation often limits its use. Trazodone is available in generic form and may be the preferred treatment in select cases of MDD depending on the patient’s unique profile. Its anxiolytic and sedative effects, for instance, may be advantageous to MDD patients with concomitant generalized anxiety disorder or insomnia [45]. Other psychotherapeutic contexts in which trazodone has been investigated include PTSD, bulimia nervosa, and adjustment disorders, among others [48–50]. More recently, a slow release preparation has been marketed that lowers plasma levels and tends to minimize sedating side effects. It should be mentioned that nefazodone is also a SARI-type ADT but largely has fallen out of use due to fear of hepatotoxicity.

Bupropion is an antidepressant that was initially released as an ADT and removed from the market due to induction of seizures in MDD patients. It was remarketed after its lower doses were noted to be safer and rereleased after the blockbuster SSRI fluoxetine went on the market. Bupropion might have been an initial blockbuster ADT as well if it had had a better safety profile. While many clinicians classify it as “atypical” because of the fact that, unlike most ADTs, it has no effect on serotonin, some prefer the specific designation of norepinephrine– dopamine reuptake inhibitor (NDRI), due to its dual mechanism that raises dopamine and norepinephrine levels instead. This gives it a unique side-effect profile characterized by no sexual dysfunction or weight gain. In fact, as it promotes weight loss, it is contraindicated in patients with eating disorders. It is also more activating with regard to insomnia and anxiety. It is frequently prescribed as part of combined ADT polypharmacy, often added to an initially, partially effective SSRI [23,51]. Meta-analysis has revealed comparable efficacy between bupropion and the SSRIs [52].

Vilazodone and vortioxetine

Vilazodone and vortioxetine are two of the most recently approved drugs for MDD. Both may be considered as “SSRI Plus” agents as their core mechanism is serotonin reuptake inhibition, but both also manipulate serotonin receptors. Vilazodone is considered SSRI-like with additionally strong 5-HT1A receptor partial agonist properties. This is somewhat similar to the mechanism of buspirone, an anxiolytic, in treating generalized anxiety disorder. Vilazodone, however, with both pre- and post-synaptic 5-HT1A agonism, is considered to be more potent in this respect. Due to the fact that this drug uses two proserotonergic mechanisms, it is sometimes termed a serotonin partial agonist reuptake inhibitor (SPARI) [53]. In animal models, there was promise of faster antidepressant effects, but this has not been replicated as yet in human trials. As a newer agent, no meta-analyses exist like those discussed previously for SSRI, SNRI, TCA, mirtazapine, or bupropion. However, it appears to have a lower risk of weight gain and sexual side effects than the SSRI or SNRI, based on various noncomparative FDA trials. A 1-year, open-label multicenter study assessing the long-term safety of vilazodone found the drug to be safe and well tolerated by MDD adults, with no clinically important changes in physical examinations, electrocardiograms, or clinical chemistries [54].

Like vilazodone, vortioxetine possesses multimodal activity and is often classified as a serotonin modulator and stimulator (SMS). In addition to its serotonin reuptake-blocking (i.e., SRI) properties, it appears to have mixed agonist–antagonist effects on a variety of serotonin receptors [55]. For example, it agonizes 5-HT1A receptors and antagonizes the 5-HT1B/D, 5-HT3, and 5-HT7 receptors, making it a unique ADT. A 2014 meta-analysis not only concluded that vortioxetine achieved significant reductions in depression scores compared with placebo in 6 of 10 RCTs, but ranked this drug above escitalopram, vilazodone, and sertraline for both efficacy and tolerability [56]. Indeed, one potential advantage of this drug over comparable ADTs is a potentially low risk of sexual side effects, weight gain, and sedation [55]. An additional advantage, as cited by McIntyre and colleagues in a recent study, may lie in its potential to improve cognitive function in adults with recurrent MDD, effects which the researchers determined to be independent of depressive symptom relief and may be due to the drug’s unique 5-HT7 receptor antagonism [57].

Of note, a large meta-analysis has been conducted to evaluate most of the modern-day antidepressants with regard to effectiveness and tolerability [58]. Specific findings here suggest that mirtazapine, escitalopram (SSRI), venlafaxine (SNRI), and sertraline (SSRI) are significantly more efficacious than duloxetine (SNRI), fluoxetine (SSRI), or paroxetine (SSRI). These mixed findings would suggest no definitive superiority of any ADT class, but rather individual drug superiority. Additionally, two SSRIs (escitalopram and sertraline) showed the best profiles of patient acceptability and thus the lowest rates of discontinuation, in support of previous data showing the SSRI class to be the best tolerated. Finally, the authors determined that sertraline for moderate-to-severe MDD may offer the best balance of effectiveness, acceptability, and cost. Of note, the newer levomilnacipran, vilazodone, and vortioxetine were not included in this meta-analysis.