TCPR: How do you choose a first-line antidepressant in major depression?

Dr. Schwartz: There are about 30 antidepressants, and they all have moderate benefits. To start, pick one that’s low risk—not a tricyclic, where there are cardiac problems and overdose fatalities to worry about, and not an MAOI, where people can die from strokes or heart attacks if they eat a tyramine-rich meal or mix the MAOI with a medication that causes serotonin syndrome. Tricyclics and MAOIs can be very effective, so I might use them if there’s no response to a first-line agent, but I’ll start with something safer—in other words, most of the antidepressants released since the 1980s.

TCPR: How do you divide those up?

Dr. Schwartz: I divide the antidepressants into roughly four categories. First, we have selective serotonin reuptake inhibitors (SSRIs), such as fluoxetine and sertraline. Next are serotonin-norepinephrine reuptake inhibitors (SNRIs), like venlafaxine and duloxetine. Third are serotonin antagonist reuptake inhibitors (SARIs), like trazodone and nefazodone, which are sedating. Fourth are norepinephrine-dopamine reuptake inhibitors (NDRIs), like bupropion, which are more activating.

TCPR: What about the newest editions: vortioxetine (Trintellix) and vilazodone (Viibryd)?

Dr. Schwartz: I’d classify them as serotonin partial agonist reuptake inhibitors (SPARIs). Like the SSRIs, they manipulate the serotonin receptors, but they do a little more. I should add that mirtazapine often gets lumped with the sedating antidepressants, but it’s really in a class of its own. It’s a serotonin receptor antagonist at a couple of places, and it enhances norepinephrine by blocking an alpha-2 receptor. 

TCPR: I’ve counted 15 post-1980 antidepressants with six receptor profiles. How do you choose among them?

Dr. Schwartz: It comes down to the patient in front of you, and side effects are usually going to guide the choice. Most people start with an SSRI, but some patients just don’t tolerate these or the closely related SNRIs. At first patients experience headache and GI distress, but that is usually manageable by slowing the titration. Over the long term they experience sexual dysfunction and weight gain, which can be a deal-breaker. Another deal-breaker is when patients feel tired or wired on SSRIs and SNRIs. 

TCPR: What does “wired” look like?

Dr. Schwartz: Agitated, anxious, and restless. You can even see overt akathisia; it is rare but possible (Koliscak LP and Makela EH, J Am Pharm Assoc 2009;49(2):e28–e38). Like other antidepressants, SSRIs can induce mania, which creates hyperactivity as well. 

TCPR: What can you use instead?

Dr. Schwartz: Before starting an antidepressant, I’ll want to know if the patient is very fatigued, or if they are agitated and can’t sleep. If they are fatigued, I’ll lean toward an activating agent like the NDRI bupropion. I may also prefer bupropion if weight gain or sexual side effects are a problem, although vortioxetine, trazodone, and mirtazapine also have low rates of sexual side effects. If they are agitated, I’ll prefer a sedating agent like trazodone or mirtazapine.

TCPR: How do SSRIs affect sleep quality?

Dr. Schwartz: SSRIs and SNRIs can disrupt sleep quality, but trazodone and mirtazapine usually improve it, possibly because they block serotonin-2A receptors (Gulec M et al, J Affect Disord 2011;134(1–3):257–265). Patients don’t sleep with an EEG on, so you have to look to other signs to pick up on this. They’ll say, “My depression didn’t make me tired before, but I’m tired now,” “I feel fuzzy and unfocused,” or, “I wake up six or seven times throughout the night.” It helps to catalog these symptoms before you start an antidepressant. Otherwise, you can end up causing side effects that look like untreated depressive symptoms. 

TCPR: You mentioned weight gain. Which antidepressants are more favorable there?

Dr. Schwartz: Bupropion. It tends to suppress appetite, and I rarely if ever see weight gain on it. Mirtazapine is probably the worst, but the other sedating antidepressants—nefazodone and trazodone—are more favorable when it comes to weight (Gill H et al, Obesity 2020;28(11):2064–2072). On average, weight gain is mild with SSRIs and SNRIs, but some people gain a lot on them, like 20–30 pounds over several years. Serotonin affects hormones involved in weight, like leptin, and may cause fat cells to grow. My guess is that some people’s genes don’t match well with that. However, the SPARIs (vortioxetine and vilazodone) seem to have an advantage here, even though they are serotonergic.

TCPR: Which antidepressants have lower rates of sexual side effects?

Dr. Schwartz: Bupropion tops that list again, but we can also add the sedating antidepressants: nefazodone, trazodone, and mirtazapine. The SPARIs (vortioxetine and vilazodone) also seem to spare the sex drive, particularly vilazodone (Reichenpfader U et al, Drug Saf 2014;37(1):19–31). 

TCPR: Do any antidepressants stand out for cognitive benefits?

Dr. Schwartz: Bupropion might help cognition. It has positive trials in ADHD and in depression, but the depression studies were not randomized (Gualtieri CT and Johnson LG, MedGenMed 2007;9(1):22). Bupropion affects norepinephrine and dopamine, which in theory are the monoamines that are most likely to improve cognition. The sedating antidepressants can go both ways. If they improve sleep quality, the patient may be a little sharper, but if they cause too much sedation they could make things worse. Then there’s vortioxetine, which tried but failed to earn FDA approval for cognitive symptoms of depression. The problem was that its benefits were not broad enough. It only changed one measure of cognition, the digit symbol substitution test, which is a measure of attention and processing speed (Huang IC et al, Int J Neuropsychopharmacol 2022;25(12):969–978). 

TCPR: Bupropion was released before the SSRIs, but the SSRIs have always been more popular. Why is that?

Dr. Schwartz: Bupropion got off to a bad start because it was pulled from the shelves for causing seizures shortly after its release. They reintroduced it a few years later with safer dosing guidelines—we don’t go beyond 450 mg/day anymore—and the extended-release forms have made it safer still (Steinert T and Fröscher W, Pharmacopsychiatry 2018;51(4):121–135). But by the time of the reintroduction, fluoxetine (Prozac) had already hit the market, and it came with an advertising blitz that bupropion never caught up with. 

TCPR: Does the patient’s age influence the choice of antidepressant?

Dr. Schwartz: Fluoxetine and escitalopram are usually first line if the patient is under age 18, as these are the only two with FDA approval in that age group. Many others tried but did not succeed. I’m also comfortable with sertraline and duloxetine, which are FDA approved in childhood OCD (sertraline) and generalized anxiety disorder (duloxetine). That reassures me of their relative safety in younger patients. If these don’t work, you’re dealing with treatment-resistant depression (TRD) and need to move up to the adult guidelines for TRD as we don’t have many data in children (Dwyer JB and Bloch MH, Curr Psychiatr 2019;18(9):26–42F).

TCPR: What about older adults?

Dr. Schwartz: For elderly folks, I start with one of the post-1980 antidepressants we just reviewed and personalize based on their symptoms and side effects, just as I’d do for general adults. The only difference is that I’ll start low and go slow, but I’ll still aim for the full dose range. Also, I’ll think more about drug interactions in older adults. Among the serotonergics, desvenlafaxine, escitalopram, and sertraline (at doses below 150 mg) have the lowest risk of drug interactions. Citalopram has a warning about QT prolongation in the elderly, so I try to avoid that one.

TCPR: Some antidepressants are more prone to withdrawal problems. How does that influence your decision?

Dr. Schwartz: Most antidepressants are associated with withdrawal phenomena, but it’s worse with the SSRIs and SNRIs, particularly the short half-life drugs: desvenlafaxine, duloxetine, venlafaxine, and paroxetine (Horowitz MA et al, CNS Drugs 2023;37(2):143–157). It’s not a lethal problem, but it is very unpleasant (like you have the flu), and these withdrawal problems are worse the longer they’ve been on the med. It’s less of an issue with a drug like fluoxetine, which has a long half-life and thus may be better suited for patients who tend to miss their dose here and there. 

TCPR: What about melancholic depression? I’m thinking about patients with early-morning awakening, worse mood in morning, low appetite, ruminative guilt, and significant psychomotor slowing or agitation. Do you prefer any antidepressants there?

Dr. Schwartz: The FDA doesn’t evaluate that subtype, but if we look outside the FDA data, I think it favors the tricyclics. If the patient can’t tolerate a tricyclic, I would go with an SNRI, which is the closest thing to a tricyclic. 

TCPR: How do you choose among the SNRIs?

Dr. Schwartz: One way they differ is in their medical risks. With duloxetine, we worry about liver toxicity, so I’m likely to avoid that in someone who drinks a lot of alcohol or takes other meds like statins that can impair the liver (Todorovi´c Vukoti´c N et al, Arch Toxicol 2021;95(3):767–789). On the other hand, duloxetine is the only SNRI that doesn’t carry a warning about raising blood pressure. Duloxetine would be a good choice for a patient with urinary incontinence (it has regulatory approval for that in other countries), but this mechanism also means it can cause urine retention in vulnerable patients. Overall, for medically compromised patients I prefer desvenlafaxine. It is the safest on the liver, and it’s less affected by drug interactions because it’s metabolized outside the P450 system through glucuronidation. 

TCPR: What about venlafaxine (Effexor) and levomilnacipran (Fetzima)?

Dr. Schwartz: Venlafaxine stands out for its withdrawal problems. Levomilnacipran doesn’t get used much, but I think it does have a niche. It is the most noradrenergic of the SNRIs, much as protriptyline is the more noradrenergic of the tricyclics. It handles a lot like bupropion, and like bupropion it can make people shaky, jittery, as well as sweaty. It’s a hard drug to tolerate, so I think you have to titrate it carefully. It’s a good option for people who are slowed down and tired. Levomilnacipran is an isomer of milnacipran, which is approved in the US for fibromyalgia (milnacipran is approved for depression in other countries).

TCPR: How do you pick and choose among the SSRIs?

Dr. Schwartz: My impression is that the SSRIs are similar in terms of response. The SNRIs are more unique, so I’m more likely to switch around in that class. If one doesn’t work, I will usually move on. In terms of ­choosing one, we talked about how their half-lives and drug interactions differ, and really those are the main things that distinguish them. Paroxetine and fluoxetine are notorious for blocking the 2D6 system. That means they are going to double or triple levels of tricyclics, which can cause fatal cardiac arrhythmias. They also raise some antipsychotics (aripiprazole, brexpiprazole, iloperidone, and risperidone), which can cause akathisia and—over the long term—tardive dyskinesia. Fluvoxamine is not a 2D6 inhibitor but does raise a lot of drug levels by inhibiting other enzymes (1A2, 2C19, 3A4, and 3A5), and it has another notch against it because it is not FDA approved for the treatment of depression, only for OCD.

TCPR: Do some antidepressants have more of a dose-dependent response?

Dr. Schwartz: For the SSRIs, most of the evidence suggests there is not a clear dose-response curve. For the SNRIs, it’s not as clear, but my experience suggests they work better as the dose goes up, and some evidence points that way as well. Vortioxetine may also have a dose-dependent response. The limitation here is that most studies are not large enough to detect differences between doses. 

TCPR: Thank you for your time, Dr. Schwartz.