Are the SSRIs really that different? If a patient doesn't respond to one will they respond to another? We sift through three decades of research to find answers.
Published On: 8/3/20
Duration: 19 minutes, 13 seconds
Tony Soprano: “Here we go, here comes the Prozac”
Welcome to the Carlat Psychiatry Podcast, keeping psychiatry honest since 2003. I’m Chris Aiken, the editor in chief of the Carlat Report. And I’m Kellie Newsome, a psychiatric NP and a dedicated reader of every issue.
The makers of SSRIs spent much of the 1990’s trying to distinguish their drug. It was a tale full of sound and fury, and signifying well, not very much. But it got people talking, and some of that chatter lives on today. In this episode we’ll look at the ways that the SSRIs tried to stand out from the crowd, and see if any of those ideas still hold water.
The clearest way they tried to stand out was through their FDA indications. Fluvoxamine was approved in OCD ─ in both children and adults. Paroxetine, perhaps recognizing that the antidepressant market was crowded, gained approval for panic disorder before depression. This Ied to the belief that paroxetine has stronger antianxiety effects than the other SSRIs. We debunked that myth last fall ─ paroxetine failed to beat other SSRIs in around a dozen head-to-head trials, and when compared to escitalopram/lexapro in generalized anxiety or depression with anxiety it consistently fared worse.
K: And there are a few reasons to avoid paroxetine. It has the worst withdrawal symptoms of any SSRI, and it’s the most likely to cause fetal malformations. It also has the strongest anticholinergic effects of the group, which may explain why it’s the only SSRI that’s been linked to dementia.
C: Others tried to distinguish themselves by their dose titration ─ sertraline and fluvoxamine have a broad dosing range so could theoretically take a while to reach the full dose, while escitalopram gets to its full 20mg pretty quickly. That’s a tenuous argument though, for example a lot of studies find that sertraline has greater antidepressant effects in the lower dose range (50-100mg) than at higher doses, possibly because higher doses simply stack on more side effects ─ apathy, sexual dysfunction, and sleep disruption ─ that are depressing in themselves.
K: One of fluoxetine’s real strengths is that it has the lowest risk of SSRI withdrawal with its long half life, which is actually not that long with fluoxetine ─ 1-4 days ─ it’s really the active metabolite norfluoxetine that lingers, with a 7-10 day half life. But this has also stirred a rumor that fluoxetine takes longer to work because it takes longer to build up in your system. Is there any truth to that?
C: Not much, and since the parent drug fluoxetine has a fairly average half-life you wouldn’t expect too much of a difference. Dr. Osser mentioned that as a possible reason why fluoxetine came up short in the latest meta-analysis, but he wasn’t sure that explained it. There was a meta-analysis in 1999 by JG Edwards that found a slightly slower onset with fluoxetine, but I could only find two head-to-head trials that looked at the question: paroxetine worked faster at 3 weeks in a 1993 study, but on the other hand the European MAOI moclobemide climbed at the same speed in a 1999 study. I wouldn’t let the speed of onset influence my decision there, but there are a few reasons to prefer fluoxetine besides the relative lack of SSRI withdrawal problems.
K: What are those?
C: 1) Eating and weight. Fluoxetine is the only SSRI approved in bulimia ─ there the dose is higher, 60mg per day. Also in some meta-analyses it had the lowest risk of weight gain.
2) Children. Fluoxetine is not the only SSRI approved in children ─ it shares that honor with escitalopram (Lexapro), but escitalopram goes down to age 12 while fluoxetine is approved to age 8. Also fluoxetine had more positive trials in children than escitalopram, where the studies were more mixed.
K: Yes but couldn’t you argue that those results are kind of random and all SSRIs probably work about the same in children?
C: Yes, but if that’s true than fluoxetine is a very lucky med indeed, because over a dozen antidepressants have been tested in children and none but fluoxetine, escitalopram, and possibly citalopram worked. And the stakes are high in this population because the warning about suicidal thinking only applies to patients under age 25, so I’d stick with fluoxetine as a first line antidepressant for children and adolescents.
K: Also there’s another advantage to the long-half life Fluoxetine would be a good choice for patients who forget to take it every day. In fact the company developed a weekly Prozac which is still on the market ─ it’s given as 90 mg a week and it’s supposed to be equivalent to 20mg a day of Prozac. What’s the story behind that?
C: Prozac weekly was released in 2001 as Prozac was going generic. I rarely use it, and not just because of cost ─ it has some problems on the clinical side that I’m going to cover in an upcoming issue.
K: Eli lily also launched a version of Prozac for women in 2000 called SaraFem, which is FDA approved for premenstrual dysphoric disorder. It’s identical to fluoxetine except it comes as a in pink/purple instead of the green/yellow capsule.
C: Sarafem was meant to take the stigma out of psychiatric treatment.
K: Some prefer to call them SRIs – and leave off the word “Selective” because so many of them have important effects on other neutrotransmitters; sertraline on dopamine; fluoxetine on norepinephrine; paroxetine’s anticholinergic effects. Nassir Ghaemi argues that escitalopram and citalopram are the only antidepressants that are truly selective for serotonin.
C: That’s all true, but it’s not clear that those differences translate into any clinical difference. You know, in the 1990’s psychiatrists were pretty skeptical about the idea that one SSRI was better than another, but most believed that the SSRIs were different. One thing I heard a lot back then was that some patients would respond to one SSRI and not another. The basis of this was a 1999 study in the Journal of Clinical Psychiatry by George Nurnberg and colleagues. I reread it today and was shocked that a study like this would create such a buzz. Here’s the conclusion:
“The SSRIs are not interchangeable, because patients who discontinue one SSRI for lack of tolerability or response can generally be treated effectively with another.”
And here’s the catch: not only was it a restrospective chart review ─ the worst kind of design ─ it didn’t even include any outcome measures to see if the patients got better. The “outcome” was how long the patient stayed on the new SSRI ─ not whether it worked. What they found was that in their residency clinic, 25% of patients who started and SSRI switched to another. They didn’t know whether the switch was due to insurance coverage, side effects, or lack of efficacy. They then noticed that these patients tended to stay on the second SSRI after the switch, but without any outcome measures it could be that they stayed on it because not because they got better but because they got tired of switching.
K: Yes patients with depression can get burned out on trying new things pretty easily. Or even if they did get better it could have been that their life changed or the illness resolved. Sounds like a pretty shady study.
C: Well it did at least find that no SSRI was the worst ─ that patients were just as likely to switch from fluoxetine as they were from paroxetine. But by overstating the conclusion it’s lead a lot of doctors to just keep switching SSRIs out of the belief that they just need to find the right switch. On some level, I think doctors want to believe that there’s a real difference between these meds. As one paper put it, citing this retrospective review as the source: “Switching from one SSRI to another appears to be effective in most cases, probably because of the significant chemical and pharmacologic differences between these agents”
K: Yea I do see doctors do switch from one SSRI to another, but most seem to believe that switching to a different class is more worthwhile.
C: That would make sense ─ at least you’re hitting different neurotransmitters ─ although with the idea that most SSRIs aren’t very selective for serotonin even that is questionable. Overall though most studies have concluded that switching antidepressants is only marginally effective, while augmenting works much better. The best data we have on switching to other classes comes from a 2008 metaanalysis of four trials by George Papakostas, Maurizio Fava, Michael Thase. They found that switching from and SSRI to a non-SSRI was more effective than switching from an SSRI to another SSRI, although keeping it within the same class was more tolerable. Still, switching out of class was only marginally better ─ you’d have to treat 22 patients to see a difference between these two strategies. Overall, switching is not the royal road to recovery.
K: Another way I’ve heard that the SSRIs is difference is in their drug interactions.
C: Yes and that one is true. Fluvoxamine (Luvox) and fluoxetine (Prozac) are potent inhibitors at several CYP enzymes, so they are going to raise a lot of drug levels. Paroxetine mainly causes a problem as an inhibitor at CYP2D6, but that’s also the enzyme that most psychotropics pass through. And then there’s sertraline. Sertraline sometimes skirts by the drug interaction radar, because it doesn’t cause a problem in the lower dose range, but once you get to 150mg a day and above it becomes a potent inhibitor at CYP2D6.
So they were all guilty until 1998 when citalopram celexa came out, although it made it’s international appearance earlier – in 1989 in Denmark. Citalopram rapidly became the drug of choice for the elderly because it didn’t have significant effects on the CYP enzymes so was pretty clear of drug interactions. It was joined on that mantle by escitalpram in 2003. We’ve suggested in the Carlat Report that all of the SSRIs interfere with bupropion’s metabolism except citalopram and escitalopram, and these two medications have some of the best clinical data when it comes to using bupropion as an augmentation agent.
K: Citalopram and escitalopram ─ we’ll call them the prams ─ may not cause many drug interactions, but they can be the victim of them. They are mainly metabolized by CYP2C19, so inhibitors at this enzyme like the PPIs that a lot of patients take for reflux can seriously raise their levels. This didn’t used to be a big deal, but now there’s a lot of concern about raising the QTc interval on these two ─ and that risk goes up as the dose goes up.
C: And that may be why there’s really no such thing as the perfect SSRI. Generally we’d prefer the prams in the medically ill and elderly because they don’t cause drug interactions, but they have the worst record when it comes to the Qtc interval, and the elderly are more vulnerable to that. In 2012 the FDA recommended not going beyond 40mg on citalopram, or 20mg in the elderly or those taking proton pump inhibitors. Now, citalopram is the worst of the two, and the only one with a black box. But let’s track back to sertraline. This is one with only mild drug interactions in the lower dose range, and that’s the dose you’d tend to use in the elderly, so it may be the better choice there. Also, sertraline has the best safety record in cardiovascular disease.
K: Sounds like we’re getting closer to the best SSRI ─ is it sertraline?
C: Well, Dr. Osser listed escitalopram and sertraline as his top pics, and they also came out on top for both efficacy and tolerability in the large network meta-analysis. Drug interactions and ─ with sertraline ─ cardiovascular safety might be other reasons to endorse these. OGBYN’s tend to prefer sertraline in pregnancy as well. But, I’ve given a few reasons to stick with good old Prozac in this podcast as well. That leaves fluvoxamine, paroxetine, and citalopram as the less favored of the bunch.
But the real message here is that we’ve spent way too much time and money mulling over this question. Time that could have been better spent supporting our patients, and money that could have directed toward more effective treatments for depression.
And now for the word of the day SPECTRAL EVIDENCE
In the legal system, there’s many kinds of evidence ─ heresay, circumstantial, and SPECTRAL.
Spectral evidence is based upon dreams and visions, and it was prominently used during the Salem witch trials of 1792. In those trials, witnesses would claim that the the spirit or spectre of the accused attacked them in their dreams. After six months of this, the court decided that spectral evidence was not a reliable method of convicting witches and it has since been banned from American jurisprudence.
The courts took another turn toward the secular in 2000 when they declared that humans are animals, and it was on this decision that Prozac lost its patent.
Got feedback? Take the podcast survey.