Trintellix claims to improve cognition in depression, but what does this effect look like in patient’s lives? We speak with Tom Gualtieri about the cognitive side of antidepressants, and show you how to use a 3 minute cognitive test that reveals a lot about how your patients are functioning.

Published On: 10/12/20

Duration: 23 minutes, 30 seconds

Article Referenced: “Keeping Up With Trintellix,” The Carlat Psychiatry Report, October 2020

Rough Transcript:

Kellie: Trintellix claims to do more than treat depression, but what exactly does it do, and how do you know if it’s working?

Dr. Aiken: Welcome to the Carlat Psychiatry Podcast, keeping psychiatry honest since 2003. I’m Chris Aiken, the editor in chief of the Carlat Psychiatry Report.

Kellie: And I’m Kellie Newsome, a psychiatric NP and a dedicated reader of every issue.

Dr. Aiken: Vortioxetine – Trintellix – has earned only one FDA approval since its release in 2013: Major depression. But the company has packed its product labeling in clever ways that aim to set this serotonergic antidepressant apart from its competitors. Most prescribing information sheets are bland affairs that rehash the results of early studies without going much beyond that. Lundbeck – the company that makes Trintellix – keeps their monograph up-to-date like a politician does to their Wikipedia page. 

Kellie: First it was the antidepressant that works in the elderly. And it’s true, the company did conduct trials with patients up to age 88. Next it was the antidepressant that worked faster than the others, and we took issue with that claim. You see, Trintellix’s product labeling states that it worked as early as 2 weeks, and Stephen Stahl – who is a speaker for Lundbeck – took that claim one step further in his popular psychopharmacology textbook. He wrote that Trintellix is unique among the antidepressants for its rapid onset – while others take about 4 weeks to work, Trintellix works as early as 2 weeks. 

Dr. Aiken: It used to be thought that antidepressants took 4 weeks to work, but Michael Posternak – who has covered Trintellix for the Carlat Report – published a paper with Mark Zimmerman in 2005 challenging that idea – they found that 60% of the antidepressant effects occurred within the first two weeks for most of the antidepressants they looked at. That work has been replicated, and it’s certainly not unique to Trintellix. Last year we asked Dr. Stahl why he endorsed Trintellix as having a specific claim for rapid onset in depression, and here’s what he wrote:

[The claim for early efficacy] was put in the 6th edition of the prescribers guide since it was in the FDA materials.  However, we took it out of the online version of the prescribers guide last year because it incorrectly implies faster onset which there is no really good evidence in favor of that.  THe two week data in my opinion are not unique.

Kellie: I don’t think he took it down though – it still says on his NEI website that Trintellix has a specific claim to work at 2 weeks.

Dr. Aiken: Well, in this month’s Carlat report Michael Posternak walks us through seven other claims for Trintellix: Generalized anxiety disorder, Age related cognitive decline, Cognitive symptoms of depression, Negative symptoms of schizophrenia, Binge eating disorder, ADHD, and lack of sexual side effects. For some of these, Trintellix actually looks pretty good, and for others it’s a dud. You can catch up on the hits and misses in the article at the carlat report.com, and in this companion podcast we’re going to focus on just one of them: Cognitive symptoms of depression.

Kellie: Didn’t they get some kind of FDA clearance for that?

Dr. Aiken: Not exactly. Lundbeck sought a second indication in 2016 for Treatment of Cognitive Dysfunction in Major Depressive Disorder. They submitted two trials to the FDA which showed that Trintellix improved cognitive functioning on a single cognitive test – the Digit Substitution Test. The FDA was not fully impressed, however, and didn’t grant the approval, but they did allow the company to add the data to its prescribing information in 2018, which effectively allows drug reps to talk about it.

Kellie: And talk about it they do. They are always passing around the Connect and Focus studies. One rep told me that the reason the FDA didn’t approve it was that they didn’t think cognitive dysfunction in depression was a separate disorder warranting separate approval.

Dr. Aiken: That’s not how I read it. The FDA was more concerned that Trintellix did not prove that it made a meaningful difference in cognition. The FDA wanted to see data that Trintellix improved the bottom line – functioning or quality of life – not a single domain of cognition that hasn’t been validated as a relevant measure of executive dysfunction in depression. 

Kellie: So what is this digit substitution test?

Dr. Aiken: It’s basically a decoder ring.

Kellie: You mean like in that movie the Christmas Story where they little boy listening to the little orphan annie radio show and they call out a series of numbers that has to translate into a message with a ring that matches numbers with letters?

Dr. Aiken: Yes. The Digit Substitution Test is very simple. You give the patient a series of 9 symbols, and each is matched with a number like they would be in a decoder ring. Then you give them a series of numbers, and they have to match as many symbols with the numbers as they can in 90 seconds. It’s a mainly a test of processing speed, but it also requires attention, motivation, and executive function to do well on it.

Kellie: But wouldn’t cognition get better with anything that treats depression?

Dr. Aiken: Probably, but in the studies they tested it against duloxetine / cymbalta and placebo. Both of the antidepressants performed similary well on symptoms of depression, and cymbalta even improved cognition some but only marginally better than placebo. For vortioxetine the improvements in cognition were more marked – compared to placebo it improved the digit sub test with a moderate effect size of 0.5.  To put that in perspective, it’s about as powerful as a shot of espresso, which also improves the DSST with an effect size in that range. On the other hand, 2 mg of lorazepam, or 4 alcoholic drinks – enough to impair driving – worsen performance on the DSST with the same effect size, 0.5, only here the effect is in the direction of harm.

Kellie: Although it’s a very simple test, performance on the digit substitution test is actually a good predictor of job functioning, and Lundbeck did do a controlled trial showing that functioning improved with Trintellix – and the effect size there was large – but I don’t think they submitted that study in time to the FDA. 

Dr. Aiken: Yes to break down the FDA’s logic a little finer:

1. They agree the Digit substitution test has some relationship to functioning, but we don’t know how much a patient needs to improve on the test to have a meaningful gain in their functioning.  More broadly, they were concerned that there’s no proof that the digit substitution test is a good reflection of the broader cognitive dysfunction seen in depression. So the drug rep you spoke with was wrong – it’s not that the FDA didn’t think cognitive symptoms of depression were unique and important enough to warrant approval – it’s actually the opposite, they thought that Trintellix’s effects were not unique and meaningful enough to warrant it.  But I don’t want to be cynical about Trintellix here – really they were saying there was a lack of evidence, not that the evidence put forth was bad. And since no other antidepressant has this kind of evidence to improve cognition, well, Trintellix may not be perfect but it’s all we got. 

Kellie: What about bupropion/wellbutrin – that one has some benefits in ADHD?

Dr. Aiken:  Yes and one study found favorable effects of bupropion – and this study was very revealing so I’m going to go into some detail about it. It was done by Tom Gualtieri who developed a computerized cognitive screening called CNS Vital Signs which is now used in research and in some practices. Dr. Gualtieri is an academic but he also has a large private practice and he tests everyone who comes through his doors with this instrument – he also tests their spouses which gives him a large sample of “normals.”  I spoke to him today:

Dr. Gualtieri tell us about the cognitive testing you do.

[Dr. Gualtieri Response]

So in this study he compared 27 patients who had recovered from depression on one of three antidepressants: SSRIs, venlafaxine, or bupropion, and he speculated that the more noradrenergic antidepressants like venlafaxine and particularly bupropion would have a better cognitive profile. He also compared them to 27 normals. The results were striking – there was no significant difference between the bupropion-treated patients and the normals on any of 5 cognitive domains – including “Psychomotor speed” which is probably the closest thing to the digit substitution test. The venalfaxine group also had better cognitive functioning, but not as good as with bupropion, and the SSRIs group came in third place.

Kellie: OK but Dr. Aiken you’re always criticizing studies that aren’t well designed – and this one was controlled with the normal sample but it wasn’t randomized and there was no placebo.

Dr. Aiken:  Yes, so it’s possible that something else accounted for this effect, like the psychiatrists were more likely to give bupropion to young healthy patients who have an active sex life. But they were all matched for age and gender and none had significant comorbidities. Dr. Gualtieri actually thinks the prescribing bias would have worked against venlafaxine and bupropion because at the time psychiatrists were more likely to use venlafaxine in severe depression and bupropion in people with cognitive problems since its benefits in ADHD were well known.

Kellie: Well it’s a shame that that data was never followed up on, because bupropion normalized all areas of cognitive functioning in Dr. Gualtieri’s study, while Trintellix only improved the digit substitution test. 

Dr. Aiken: Yes bupropion has never gone head to head with Trintellix, but other medications have and their results are pretty consistent with what Dr. Gualtieri found. The SSRIs do not improve performance on the digit substitution test; surprisingly neither do the MAOIs or tricyclics – even though the tricyclics have noradrenergic effects they also have anticholinergic and other effects that might block that benefit. Venlafaxine was not tested but another SNRI – duoloxetine/cymbalta –was the only antidepressant to improve the DST besides vortioxetine, and it didn’t improve it as well as vortioxetine did.

Dr. Gualtieri – how important is the digit substitution test – what difference does it make in people’s lives?

[Dr. Gualtieri Response]

Dr. Aiken: One criticism of vortioxetine’s data is that it only made a small difference on the digit substitution test. When we remove the placebo effect, people’s score improved about 4 points, which means that instead of decoding 40 digits they decoded 44. The effect size for that difference is in the small range, from 0.3-0.5 depending on the study, which is about the effect size you’d get if you drank a shot of espresso before the test.

Dr. Gualtieri what difference does that make in everyday life?

[Dr. Gualtieri Response]

Dr. Aiken: How important is cognition when selecting an antidepressant?

[Dr. Gualtieri Response]

Dr. Aiken: Dr. Gualtieri you test every patient you see with the digit substitution test and a whole 30 minute cognitive battery. I’m wondering how patients respond to all that testing?

[Dr. Gualtieri Response]

Dr. Aiken: So sometimes the test is not right?

[Dr. Gualtieri Response]

The digit symbol substitution test has been around for over a hundred years, and we’ve included a link to it in our show notes – or just go to moodtreatmentcenter.com/measurement and you’ll find it there.  That link also has the normative values for different populations, which is informative, thanks to our intern Kaira Raeiteri for preparing that. In children the scores get better with age all the way up to age 18. Performance peaks around age 25 and slowly declines after that and the decline becomes sharper in the late 40’s . It falls , from a mean of 78 in the early 20’s to a mean of 42 at age 80 – and that’s also the average score for an adult with depression – 42, while for schizophrenia it’s just a little bit lower at 38. Women tend to perform better than men, but interestingly there is no effect of education – this is one of the few cognitive tests where people of all educational levels perform about the same. Same thing with intelligence.

Kellie: Here’s how you might use the test in practice. Print it out and have the patient do it before they start a new medication, and then again 4 weeks later. It only takes 3 minutes to do so you can do it in your office, but if you don’t have time you can ask the patient to do it at home, using a stop watch – just make sure they do it at the same time of day in the same place on the before and after tests.  You don’t just have to use it before starting vortioxetine – because any antidepressant is likely to improve the performance just by treating depression – remember vortioxetine’s claim to fame is that it improved performance independent of its mood benefts and better than other antidepressants did.

I find that testing like this helps engage patients in their treatment. They are more likely to stick with a medication if they see that it makes a difference on the test. And they are more likely to stick with behavioral interventions for cognition like exercise, mediteranean diet, and of course sleep. A night without sleep will drop your digit substitution score by about 14 points – which is about 3 times worse than the effect of lorazepam 2mg or drinking enough to tip you over the legal blood alcohol limit – around 3-4 drinks.

Dr. Aiken: Tom Gualtieri is the Medical Director of North Carolina Neuropsychiatry, a practice which grew out of his academic work at UNC Chapel Hill. He has authored over 120 peer-reviewed medical publications and several books, most recently The OCD of Everyday life.

Kellie: And now for the word of the day….  Memory hallucinations

Dr. Aiken: Memory hallucinations are false memories conjured up retrospectively. They are seen in patients with chronic psychotic disorder and often have a fantastical element, such as a patient who tells you they were abducted by a band of gypsy’s as an child or that their parents were part of a satanic cult.  Patients may recall prophetic visions that later came true. These inserted memories often have paranoid themes, or these of guilt like recalling that they committed a crime which they didn’t. Sometimes, of course, the fantastical turns out to be true.  False memories were first described in the 1800’s by hypnotists who implanted them into the memories of normal people under a hypnotic spell. The French hypnotist Hippolyte Bernheim called these retroactive hallucinations, though they are not actually of a psychotic nature. In 1886 he described a case where he had suggested to a woman in a trance that she had fallen and hurt her nose while getting out of bed the night before. When questioned later, she recounted the fall and insisted that she had a clear memory it.

Join us next week for : An Interview with Ruta Nonacs on Menopause and Mood

Kellie: You can read the full articles online including this month’s review Keeping up with Trintellix by Michael Posternak which covers new research on Trintellix in 7 different conditions.U

Dr. Aiken: The Carlat Report is one of the few CME publications that depends entirely on subscribers. Thank you for helping us stay free of commercial support.

Got feedback? Take the podcast survey.