Duration: 19 minutes, 35 seconds

KELLIE NEWSOME: Today we count down the top 10 research updates of the past year. Welcome to the Carlat Psychiatry Podcast, keeping psychiatry honest since 2003.

CHRIS AIKEN: I’m Chris Aiken, the editor-in-chief of the Carlat Report. 

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

CHRIS AIKEN: Every September, I present the top 10 research findings of the past year at the North Carolina Psychiatric Association’s annual meeting, and I’m going to count them down here as this year’s talk was canceled by Hurricane Helene. But I am going to count them down here, from 10 to 1. The trials are listed in no particular order, and we’re going to start with two that have changed how I think about treatment-resistant depression.

These two trials compared several popular strategies for treatment-resistant depression: transcranial magnetic stimulation (TMS), switching antidepressants, or augmenting the antidepressant with either a tricyclic or aripiprazole. In each study, TMS won the day, but there are subtleties here that will turn down that enthusiasm a few notches.

Both of these trials enrolled patients with true treatment resistance, having failed at least 2 antidepressants. Both lasted 8 weeks, and both were funded by non-profits. Both were randomized controlled trials, but neither of them involved a placebo arm, and neither were double-blinded. That means the patients and their doctors were aware of what treatments were being used – it is, after all, hard to blind people to the use of TMS but if they had more funding they could have arranged a sham TMS arm. The raters, at least, were blind to the treatments. Let’s look at these two trials in more detail.

The largest trial is the ASCERTAIN-TRD trial, a multi-site study involving 278 patients with TRD who were randomized to rTMS which was added to their current antidepressant, augmentation with aripiprazole (with a mean dose of 9 mg), or switch to an SNRI (either venlafaxine mean dose 191 mg or duloxetine mean dose 98 mg). Although there was no placebo arm, we know that after two failed antidepressants switching is no better than placebo, so the real question here is which of these were better than switching? 

Only rTMS surpassed switching, with a meaningful difference of 4 points on the MADRS. That begs the question – why didn’t aripiprazole work here when it has passed in so many other trials? And unlike many other antipsychotics, aripiprazole passed in trials of true treatment resistance – at least 2 failed antidepressants. The best guess is that the study was underpowered, for a 3-arm study they may have needed more than 278 patients to detect a difference. And the reason some of it was underpowered, is that some of it was cut short by COVID.

The ASCERTAIN study was sponsored by the Patient-Centered Outcomes Research Institute and was published by George Papakostas and colleagues in Molecular Psychiatry, March 2024.

Our second trial is a smaller one, enrolling 89 patients with treatment-resistant depression. This study is a little unclear unless you understand its purpose, which was to test whether TMS should be moved up in the Dutch treatment guidelines for major depression. So the patients were randomized to either add on TMS or to take the next step in the Dutch guidelines for TRD, which is to either switch to a tricyclic or augment with lithium or an atypical antipsychotic. In the trial, only 13% ended up augmenting, the majority of them switched to a tricyclic, so I think about this Dutch arm as if they all switched to a tricyclic. 

After 8 weeks, rTMS did better than the Dutch switch and augment on all outcomes, and the difference was substantial, with an effect size of 0.8. The study was published by Iris Dalhuisen and colleagues in the American Journal of Psychiatry in September 2024.

Together, what these studies tell us is that we should be thinking about TMS when patients haven’t recovered after two antidepressant trials. The limitation here is the lack of blinding – which means the fancy technology and close clinical attention of the TMS group may have swayed the data. I will grant that – and I think it’s more the close attention and the fact that you need to get out of bed every day and go to treatment that ups the placebo effect with TMS. But I don’t think that explains all the changes here. We already know that TMS has a large effect size in general–around 1.3. Where the effect is even larger with ECT, around 2.1. So the only treatments with a big umph in treatment-resistant depression are TMS, ECT, and actually ketamine, but ketamine’s benefits are short-lived. So I would place more weight on TMS or ECT. ECT is also short-lived, but at least we have evidence-based strategies to keep it going, like psychotherapy, the combination of nortriptyline and lithium, or monthly maintenance ECT. After a 6-week course of TMS, most patients stay well, and if they don’t the treatment can be delivered weekly or monthly to keep them well. 

The main drawback to TMS is the cost, around 6,000 to 12,000 for the entire course, but insurance usually covers it. There is a rare risk of seizures and common side effect of headache or jaw tension. And if you send your patient to TMS, here’s a tip: The treatment works better if they do it in an active, positive frame of mind. How do you get a depressed person in a positive frame of mind without invalidating them? I ask patients to walk in nature, read a favorite spiritual or literary text (like a poem or book of prayer), listen to engaging music, or do some active problem-solving like a crossword puzzle before they sit down for the TMS session.

KELLIE NEWSOME: Our next study changed how I think about benzo withdrawal. It’s a retrospective review of over 200,000 patients on long-term benzo from a commercial insurance database. They excluded patients who were taking the benzo for cancer- or hospice-related anxiety, seizures, and anyone with a history of drug overdose. Next, they divided the sample into those who either came off or stayed on the benzo and then looked at their mortality outcomes over the next year. Surprisingly, mortality rates were 1.6 times higher in those who came off the benzo compared to those who stayed on. This was not a randomized controlled trial, so it’s possible that what is really going on is that doctors decided to taper the benzos off in sicker patients, and sicker patients naturally have higher mortality rates. But the results held up even after adjusting for major confounders like age, sex, race, location, benzo-dose, neuropsych meds, and comorbidities like pain, insomnia, bipolar, psychosis, and addiction. Another surprise in this study is that the same pattern – that one-and-a-half-fold higher mortality rate – held true regardless of the patients age and whether they were taking an opioid.

CHRIS AIKEN: In my own practice, it’s those two patient groups where I’ve been most aggressive about tapering down benzos, the elderly and those on opioids. But this study doesn’t encourage me that I’m saving any lives here. Now, if you're still doubting the results of this study, as you have good right to do as it was not randomized. There was a secondary analysis that draws a more plausible connection to benzo withdrawl as the cause of the mortality rates here. This secondary analysis looked at the kind of risks that we worry about with benzos: Overdose attempts, suicide attempts, and visits to the ED. All of these were higher in those who came off benzos as well, and that points to a possible reason for the higher risk here. The authors speculate that some of the increased mortality might have been due to people turning to street drugs – including street benzos – as their prescribed benzo was discontinued. But, we don't have data on that, so we just don't know.

KELLIE NEWSOME: What this study tells me is something we could have learned by listening closely to patients. For many, benzo withdrawal is a miserable experience. We hear that in practice, but it’s hard to tell if the patient is complaining loudly because they miss the rewarding effects of these controlled substances or because they are really having a bad withdrawal. You can’t tell, but based on this data I would err on the side of compassion. 

CHRIS AIKEN: It is my guess for some that the absolute misery of benzo withdrawal may have driven them to desperate, dangerous, or at least unhealthy behaviors that contributed to this mortality rate. In psychiatry, we deal with subjective symptoms, but subjective symptoms have serious objective consequences, like Akathisia, for example, raises the risk of suicide. Benzodiazepines also raise the risk of suicide long-term, or at least they seem to from non-randomized trials. What this new non-randomized trial tells us is that as bad as long-term benzo use may be, and yes it does have problems with it, withdrawal maybe even worse. We saw a similar pattern in a large VA study of prescription opioid withdrawal, where patients were taken off opioids or kept on them and the ones who came off had worse mortality rates and suicidality outcomes. Here is another way this trial has changed me, when we asses a patient we think about their diagnosis and the risks and benefits of treatment, with this study I’d add a fourth leg to that decision tree – the risk of medication withdrawal. We all inherit patients who are on inappropriate medications – whether multiple antipsychotics or high doses of stimulants and benzos, and we need to weigh the risk of withdrawal as much as continuation. This study was by Donovan Maust and colleagues from December 2023’s JAMA open network.

KELLIE NEWSOME: Our next study gives us some reassurance about long-term use of benzos. It looked at the entire population of Denmark over a 20-year period, over 4 million people. In that time, 22% of them started a benzo, and this study looked at whether starting a benzo led to escalating doses or chronic use. They included z-hypnotic sin the benzo group.

It did not. Long-term use was rare. Among the benzo users, only 15% stayed on for more than a year, 5% for more than 3 years, and 3% for more than 7 years. Long-term use was more common with z-hypnotics than benzos, and – not surprisingly – more common in those with a history of a substance use disorder.

Dose escalation was also rare. Instead, doses tended to decline over time. The definition of dose escalation is useful to bear in mind for your own practice: They defined it as going beyond 40 mg diazepam equivalents for adults, and 20mg diazepam equivalents for those 65 and older. After 3 years of benzo use, only 7% escalated the dose, which if we look at all who started a benzo that’s only 0.4% of the sample. As with chronic use, dose escalation was more common with z-hypnotics and those with a substance use history.

Now, one question is whether these encouraging results will generalize outside of Denmark, a country that has taken steps to limit benzodiazepine use since the 1980s. We don’t know, but the results are similar to 2003 study of a New Jersey Medicaid population.

The bottom line: In general medical practice, dose escalation and chronic use is rare with benzodiazepines, but you ought to watch out for it when your patient has a history of substance use disorder. You don’t have to taper all your patients off a benzo but try to keep those doses below 40mg diazepam equivalents in adults and 20 mg in the elderly. Use a benzo dose converter from your favorite textbook or website to figure those out.

CHRIS AIKEN: And if you do decide to taper the benzo, the American Society of Addiction Medicine recently published the first guidelines for doing so. You’ll find the link to that – and to all the studies – in the script for this episode at the Carlat website. Just google Carlat Psychiatry Podcast. Here are some of the highlights from the new guidelines: They recommend seeing patients who are on benzodiazepines at least every 3 months, whether you are tapering them or not. And you should consider a taper in patients who are at risk for falls, car accidents, cognitive impairment, or overdose. In my own practice, I’ll attempt a taper in any older adult who has had a recent fall, regardless of the cause of that fall. Other groups where tapering is recommended include the elderly, pregnant women, and those with a history of substance use disorders, or who are taking opioids.

What? At this point you might be thinking – didn’t that new study find higher mortality rates in those groups with the benzo taper? Well yes, but there is other study showing major risks in those groups, like that benzos raise the risk of opioid overdose approximately 4-fold when the two are taken together, and I’ll caution that conflating results are not uncommon in psychiatry, we still have to do our best in the face of ambiguity, and these new guidelines did not include this brand new study on benzo withdrawal and mortality. 

Finally, how to taper. This is more of an art than a science. As a starting place, the guidelines recommend tapering by 5-25% every 2-4 weeks. You can usually lower the dose faster, by 25%, in the beginning, and slower as you get down to the bottom of the barrel. Here’s a helpful rule of thumb. To avoid abandoning the taper, I will tell the patient that we can slow it down if things are difficult, but that we can’t raise it back to an earlier dose. I also will not  start the taper if the patient seems unsure, like if they ask repeated “what if” questions like “What if I can’t sleep… what if I can’t function” and I’m not able to reassure them. I also recommend psychotherapy strongly for anyone undergoing a benzo taper. CBT has the best evidence to help benzo withdrawal, while medications have largely proven unsuccessful. If they don’t start therapy, I’ll recommend the Stopping Anxiety Medication Workbook by Michael Otto. 

However you do it, make sure that a strong working alliance is in place before tapering the benzo – that makes all the difference in success. If that’s not in place, patients may drop out, get argumentative, or even downright threatening. If things turn sour, the guidelines have a useful section aptly titled, “When a shared decision cannot be reached with the patient.”

KELLIE NEWSOME: You can find the slides for this podcast summarizing all of the top 10 studies at chrisaikenmd.com/speakerrequest. Join us next week for updates on lithium, lumateperone, and lavender. 

Can’t wait till then? Get more updates in the September edition of Carlat Psychiatry News – search for it on YouTube. We cover new guidelines on benzo tapering and new studies of brexpiprazole in PTSD, antidepressants in bipolar, oral ketamine, CBD oil, e-Cigs, and fraudulent stimulant scripts.

The Carlat CME Institute is accredited by the ACCME to provide continuing medical education for physicians. Carlat CME Institute maintains responsibility for this program and its content. Carlat CME Institute designates this enduring material educational activity for a maximum of one quarter (.25) AMA PRA Category 1 Credits^TM. Physicians or psychologists should claim credit commensurate only with the extent of their participation in the activity.