Printed from:
Marooned: Only one choice


Journal of Practical Psychiatry and Behavioral Health, March 1998, 110-114

The questioner finished by asking: "So which one would you choose?" The question came at the end of a lecture. The questioner had set the stage by posing the following hypothetical scenario: I was stranded on a deserted island and could have only one medication, an antidepressant, available. Given those facts, which antidepressant would I choose? Before I give my answer, I’ll lay out some necessary background to help you understand the rationale behind it. In discussing this background information, I will present a way of conceptualizing and organizing available antidepressants into eight classes based on their mechanisms of action and will explain how this classificatory system can help the prescriber:

Rational Drug Development

Neuropsychopharmacology has entered an exciting new era because of the ability to rationally develop new medications to treat psychiatric/neurologic illnesses. Rational drug development involves selecting a target of interest and then determining the molecular structure needed to bind to that target but not to other targets that could mediate tolerability or safety problems.

Major depression is the first therapeutic area in psychiatry that has benefited from such an approach to psychiatric drug development-the result has been an explosion of antidepressant options. In fact, this approach has been so successful that a new antidepressant has been approved for marketing in the United States every year for the past 10 years. While this explosion in options has benefited patients, it has meant that prescribers must remember and weigh more facts every time they consider an antidepressant for a patient. In earlier columns (November, 1995; January and September, 1997), I presented a system for remembering the potential for antidepressants to interact pharmacokinetically with other medications. In this column, I present a classificatory system to help conceptualize and group the available agents pharmacodynamically. This classification will help the prescriber anticipate the effects of these different antidepressants, think about sequential treatment options for the patient who has not benefited from a trial of a specific type of antidepressant, and anticipate the kinds of pharmacodynamically mediated drug-drug interactions that may result from the use of a specific type of antidepressant in combinations with other medications.

A Mechanistically Based Classificatory System

In the past, there was no uniform way of classifying and referring to antidepressants. Some, such as tricyclic antidepressants (TCAs), were classified according to their structure. Others, such as monoamine oxidase inhibitors (MAOIs), were classified according to their principal pharmacological effect, which was also the presumed mechanism of action responsible for their antidepressant efficacy. Although this classificatory system was not satisfying, it was the best we had for many years. It caused no great problems because the only antidepressants available belonged to one of these two classes.

Rational development made this old haphazard classificatory system passé and led to the development of a new classificatory system based on the pharmacology of the various agents (Table 1). The advantage of this approach is that it facilitates understanding of the drugs’ physiological effects and their potential pharmacodynamically mediated drug-drug interactions.

Table 1.  Classification of Antidepressants by Putative Mechanism(s) of Action Responsible for Their Antidepressant Efficacya
  • Serotonin and norepinephrine reuptake inhibition plus effects on multiple receptors and fast sodium channels (e.g., amitriptyline, imipramine)

  • Serotonin selective reuptake inhibitors (fluoxetine, fluvoxamine, paroxetine, sertraline)

  • Norepinephrine selective reuptake inhibition (e.g., desipramine)

  • Serotonin and norepinephrine reuptake inhibition (e.g., venlafaxine)

  • Serotonin-2A (5-HT-2A) receptor blockade and serotonin uptake inhibition (e.g., nefazodone)b

  • Serotonin (5-HT-2A and 2C) and norepinephrine alpha-2 receptor blockade (e.g., mirtazapine)b

  • Dopamine and norepinephrine reuptake inhibition (e.g., bupropion)

  • Monoamine oxidase inhibition (e.g., tranylcypromine, phenelzine)c

aThe presumptive mechanism of action for each drug is based on the preclinical pharmacology of the drug and the fact that it reaches sufficient concentration in vivo to affect this site of action given its in vitro potency.

bBoth nefazodone and mirtazapine also have other mechanisms of action that are engaged at concentrations which occur under clinically relevant dosing guidelines.

cOnly irreversible and nonselective monoamine oxidase inhibitors (MAOIs) are available in the United States, but selective and reversible MAOIs are marketed elsewhere in the world.

The ability to classify antidepressants in this fashion is based on the rational drug development process that was used to discover the newer antidepressants. In this process, the tertiary amine TCAs (i.e., amitriptyline, doxepin, imipramine, trimipramine) were used as the blueprint for what newer antidepressants should and should not do. In this discussion, I will use amitriptyline as the prototypic tertiary amine TCA. This is an appropriate choice since amitriptyline was for many years the most widely used antidepressant in the United States and throughout the world. In fact, it is still widely used to treat a number of different conditions (e.g., chronic pain, fibromyalgia). Nevertheless, the use of amitriptyline and other tertiary amine TCAs is fraught with difficulties because of their multiple mechanisms of action. To understand the pharmacology of amitriptyline, imagine the number of drugs with a single mechanism of action one would have to take to achieve all the actions that can be achieved with amitriptyline alone (Table 2). At low doses (or concentrations), amitriptyline blocks histamine receptors. At higher doses, amitriptyline sequentially blocks the other sites outlined in Table 2.

Table 2.  Amitriptyline:   Polypharmacy in a Single Pill
Drug Sharing Mechanism of
Action with Amitriptyline










Histamine-1 receptor blockade

Histamine-2 receptor blockade

Acetylcholine receptor blockade

Norepinephrine uptake inhibition

Serotonin uptake inhibition

5-HT2 receptor blockage

NE-alpha-1 receptor blockade

NE-alpha-2 receptor blockade

Direct membrane stabilization

Note: The multiple actions of amitriptyline are listed in descending order of potency (i.e., histamine-1 receptor blockade is the most potent, whereas direct membrane stabilization is the least.)

The newer antidepressants were designed to reproduce some but not all of the effects of amitriptyline. Specifically, the newer agents were designed to produce the effects of amitriptyline that were thought to be responsible for its antidepressant efficacy (e.g., serotonin and norepinephrine uptake inhibition) and to avoid effects that were likely to be responsible only for adverse effects rather than antidepressant efficacy (e.g., cholinergic and alpha-1 adrenergic receptor blockade).

Table 3 lists six of the eight classes of antidepressants that are now available based on their effects on specific mechanisms of action. The first column of this table lists the specific neural mechanisms of action shared by one or more than one class of antidepressants in the order that these targets are affected by amitriptyline (i.e., with the most potent mechanism, histamine receptor blockade, listed at the top and the least potent, inhibition of sodium [Na+] fast channels, listed at the bottom). In Table 3, a "yes" indicates when the drug, at usual therapeutic concentrations, would be expected to affect that target to a physiologically significant degree. A "no" indicates when this would not be expected under usual therapeutic dosing conditions.

Table 3.  A Comparison of the Mechanisms of Action of Antidepressants Representative
of Six of the Eight Classes Outlined in Table 1.a
Mechanism of Actionb Amitriptyline Desipramine Sertraline Venlafaxine Nefazodone Mirtazapine
Histamine-1 receptor Yes No No No No Yes
Acetylcholine receptor blockade Yes No No No No No
Norepinephrine (NE) uptake inhibition Yes Yes No Yes No No
Serotonin 5-HT2A receptor blockade Yes No No No Yes Yes
Alpha-1 NE receptor blockade Yes No No No Yes No
5-HT uptake inhibition Yes No Yes Yes Yes No
5-HT2C receptor blockade No No No No No Yes
5-HT3 receptor blockade No No No No No Yes
Alpha-2 NE receptor blockade No No No No No Yes
Inhibition of fastc sodium channels No No No No No No
a Monoamine oxidase inhibitors (MAOIs) and bupropion are not shown. MAOIs are not shown since they do not directly share any mechanism of action with any other class of antidepressant, although they affect dopamine, norepinephrine, and serotonin neurotransmission via their effects on monoamine oxidase. Bupropion is not shown because its mechanism of action remains unclear. Its most potent action is the inhibition of dopamine and norepinephrine uptake but its effect on these sites of action in vitro is so weak as to raise questions as to whether these actions are clinically relevant.
The effects of these various antidepressants are listed using a binary (yes/no) approach for simplicity and clinical relevance. The issue for clinician and patient is whether the effect is expected under usual dosing conditions. A "yes" means that usual dosing of the drug for therapeutic purposes typically achieves concentrations in the usual patient that should engage a specific target to a physiologically/clinically significant extent (given the in vitro affinity of the drug for that target). If the typical concentration of a drug is ten or more times less than the drug’s affinity for a specific target, then it is unlikely to affect that target to a physiologically/clinically significant extent under usual dosing conditions. For example, under usual dosing conditions, amitriptyline achieves concentrations that engage the norepinephrine uptake pump. At such concentrations, it also substantially blocks histamine-1 and muscarinic acetylcholine receptors since it has even more affinity for those targets than it does for the norepinephrine uptake pump. Since the binding affinity of amitriptyline for the serotonin-2A and alpha-1 norepinephrine receptors and the serotonin uptake pump are within an order of magnitude of its affinity for the norepinephrine uptake pump, amitriptyline at usual therapeutic concentrations will also affect those targets. On the other hand, amitriptyline will not typically affect Na+ fast channels at usual therapeutic concentrations, but can do so in an overdose situation. This accounts for the narrow therapeutic index of the TCAs and is the reason therapeutic drug monitoring to detect unusually slow clearance is a standard of care when using such drugs (see my May, 1996 column).
The binding affinities of all drugs listed in the table (except mirtazapine) are based on the work of Cusak et al.1 and Bolden-Watson and Richelson.2 Information on the binding affinities of mirtazapine (including affinities for 5-HT2A and 5-HT3 receptors) is based on the work of deBoer et al.3 Although the publications by the other group did not include values for the 5-HT2A and 5-HT3 receptors for the other antidepressants, Elliot Richelson of the Mayo Clinic in Jacksonville, FL (personal communication) confirmed that the other antidepressants would be unlikely to affect these receptors under usual dosing conditions.

As indicated in Table 3, antidepressants are now available that selectively block the uptake of either norepinephrine (e.g., norepinephrine selective reuptake inhibitors, NSRIs) or serotonin (serotonin selective reuptake inhibitors, SSRIs). We also have antidepressants that have more than one mechanism of action over their clinically relevant dosing range (e.g., mirtazapine, nefazodone, venlafaxine).

As can be readily seen, the newer antidepressants reproduce some but not all the effects of amitriptyline. For example, the SSRIs and amitriptyline share the ability to block the neural uptake of serotonin but do not share the ability to block the alpha-1 adrenergic receptor. Because they avoid that effect, SSRIs do not cause the orthostatic hypotension and resultant falls that can plague the use of amitriptyline

Tertiary amine TCAs such as amitriptyline inhibit Na+ fast channels at concentrations that are only 5 times higher than those needed to block the uptake of norepinephrine and serotonin. While the inhibition of the uptake pumps for serotonin and norepinephrine is the mechanism of action believed to be responsible for the antidepressant efficacy of the TCAs, the inhibition of Na+ fast channels is responsible for their cardiotoxicity and their high lethality when taken as an acute overdose. Therefore, one goal of rational drug development was to widen the therapeutic index of newer antidepressants by developing antidepressants which did not inhibit Na+ fast channels.

Table 4 lists the clinical consequences that can occur as a result of antagonism of the various sites of action listed in Table 3. The prescriber can use Tables 3 and 4 in combination to determine what effects a specific type of antidepressant would be expected to produce in the usual patient on such a medication alone at usual therapeutic doses. The prescriber can also use these two tables in combination to anticipate what type of pharmacodynamically mediated drug-drug interactions are likely to occur when a specific type of antidepressant is combined with other agents. For example, all drugs that block the serotonin uptake pump can interact with monoamine oxidase inhibitors to cause the serotonin syndrome. Similarly, all antidepressants that block the histamine receptor centrally will potentiate the cognitive-motor effects of alcohol, and the propensity of all antidepressants that block the alpha-1 adrenergic receptor to cause orthostatic hypotension is potentiated when they are coprescribed with other antihypertensive medications. Thus, the use of these two tables to anticipate pharmacodynamic drug-drug interactions is analogous to the use of the two tables in my November, 1995 column to anticipate pharmacokinetic drug-drug interactions mediated via cytochrome P450 enzymes.

The prescriber can also use this table to guide drug selection when a patient cannot tolerate the first antidepressant tried or the first antidepressant does not work. In such situations, the most rational approach is to try an antidepressant with a different presumed mechanism of action mediating its antidepressant efficacy rather than switching around within a class. As I discussed in my July, 1997 column, only open label studies have been conducted to test switching among members of the class of SSRIs in the event that an adequate trial of one SSRI has not worked. As I discussed in that column, the results of such open label studies are highly suspect and, until adequate double blind studies are done, it would be most reasonable to conclude that the overlap spectrum of antidepressant activity among the various SSRIs is so similar that switching a patient who has not benefited from one SSRI to another SSRI is not likely to produce meaningful benefit.

Table 4.   Sites of Action and Clinical and Physiological Consequences of Their Blockade or Antagonism
Site of Action Consequences of Blockade
Histamine-1 receptor Sedation, antipruritic effect
Muscarinic acetylcholine receptor Dry mouth, constipation, sinus tachycardia, memory impairment
Norepinephrine (NE) uptake pump Antidepressant efficacy, increased blood pressure, tremors, diaphoresis
Serotonin (5HT2) uptake pump Antidepressant efficacy, nausea, loose stools, insomnia, anorgasmia
5-HT2A receptor Antidepressant efficacy, increased REM sleep, anti-anxiety efficacy, anti-EPS
Alpha-1 NE receptor Orthostatic hypotension, sedation
5-HT2C receptor Anti-anxiety efficacy, increased appetite, decreased motor restlessness
5-HT3C receptor Antinauseant
Alpha-2 NE receptor Antidepressant efficacy, arousal, increased libido
Sodium fast channels Delayed repolarization leading to arrhythmias, seizures, delirium

What About the Desert Island and the Antidepressant?

Looking at Tables 3 and 4, the answer is easy. If I were on a desert island and could have only a single medication (an antidepressant) available, my choice would be amitriptyline. My rationale is as follows: A low dose of amitriptyline could be taken for its antihistamine properties if I was exposed to the tropical variant of poison ivy while on the island. A somewhat higher dose of amitriptyline could be taken for its anticholinergic effect, if I inadvertently ate "bad" food and got diarrhea. A somewhat higher dose could be taken for its alpha-1 adrenergic antagonism if I developed hypertension-perhaps from too much salt. Obviously, one could also take it if a depressive episode developed. I could even use amitriptyline, because of its effect on NA+ fast channels, to treat certain types of cardiac arrhythmias. However, too high a dose would carry the risk of a fatal arrhythmia for the same pharmacological reason.

Given the hypothetical scenario set up by the questioner I mentioned at the beginning of this column, amitriptyline would obviously be my choice because it is in essence several medications in a single molecule. None of the other antidepressants shown in Table 3 have the versatility of amitriptyline. In fact, these other antidepressants were rationally developed to avoid just such versatility because they were not intended to be used by people marooned on an island. The reason amitriptyline would be the choice in the hypothetical scenario is precisely the same reason that we tend to avoid it when treating a patient with an uncomplicated form of major depression in conventional practice. Since these patients are not marooned with only a single medication available to treat them, the prescriber can use other medications to selectively achieve only desired effects while avoiding undesirable effects. Thus, we could use a newer antidepressant that avoids histamine receptor blockade to treat our patient and add an antihistamine should the patient develop a case of poison ivy and then stop it when the rash resolves. With amitriptyline, we cannot separate its antidepressant effect from its antihistaminic effect because they both come in the same molecule. Thus, what may be a short-term advantage for amitriptyline often becomes a long-term disadvantage for the patient.

Some readers might wonder why we still talk about multiple targeted medications like amitriptyline and other tertiary amine TCAs. However, they might be surprised to realize that several of the "newer" atypical antipsychotic medications are essentially multiple targeted medications. For example, olanzapine is to haloperidol as amitriptyline is to fluoxetine or any other SSRI. Thus, newer antidepressants are marketed as being "selective" while many of the recently introduced antipsychotics are marketed as being "atypical."

The point is that multiple targeted medications are neither inherently good or bad-rather the physician must simply realize that such drugs will produce more varied kinds of effects, typically in a dose dependent fashion, than will more selective medications. Those effects may be good or bad depending on the clinical circumstances, but clearly the more effects a medication has, whether it is amitriptyline or olanzapine, the more ways it can interact pharmacodynamically with other coprescribed medications. By keeping these principles in mind and by knowing which mechanisms of action a drug is likely to engage under usual dosing conditions, the physician will be better able to anticipate what effects that drug will be likely to produce in his or her patients when used alone or in combination with other medications.


  1. Cusack B, Nelson A, Richelson E, Binding of antidepressants to human brain receptors: focus on newer generation compounds. Psychopharmacology. 1994;114:559-565
  2. Bolden-Watson C, Richelson E, Blockade by newly-developed antidepressants of biogenic amine uptake into rat brain synaptosomes. Life Sci. 1993;52:1023-1029
  3. DeBoer TH, Ruigt GSF, Berendsen HHG, The alpha 2 - selective adrenoceptor antagonist Org 3770 (mirtazapine, Remeron) enhances noradrenergic and serotonergic transmission. Human Psychopharmacology 1995;10:S107-S118

Suggested Readings