Do Benzos Treat Depression?

Around 50 controlled trials have tested benzodiazepines in depression, and the results are surprising. In this episode we look at the controversies surrounding this research and what it means for practice.

Listen above or open in your podcast store

From Barbiturates to Benzos

“Nothing is more tormenting to a melancholic than to lie awake for several hours in the early morning, alone with his morbid thoughts and his intense desire for the peace of sleep. Judicious prescription of the proper dosage of the right drug is most rewarding to such a patient.”That was from Frank Ayd’s 1961 textbook Recognizing the Depressed Patient, and the “right drug” he was talking about was the barbiturates. Soon after the book’s release, barbiturates were replaced with the benzodiazepines, and the benzos were later supplanted by the z-hypnotics, at least for sleep. It sounds like progress, but we aren’t too far from Dr. Ayd’s world. All of these drugs are gaba-ergic, and their main difference is in safety, not efficacy. In fact, the older versions are arguably more potent than the newer gaba-ergics we use today. But they were also more toxic and more addictive. Barbiturates were fatal in overdose on their own; while benzos mainly cause overdose deaths when combined with other drugs like opioids or alcohol. The Z-hypnotics have been with us since the late 1980’s, and we’re still debating whether they lead to dependence or significant overdose problems.

KELLIE NEWSOME: And through all these changes, the gaba-ergic medications have remained common player in the therapy of depression since the 1950’s. They’ve always occupied an awkward place. Practitioners use them but don’t quite accept them; regulators discourage them; and patients appreciate them, sometimes too much. In this two-part series we’ll cover this uncomfortable history, starting with the barbiturates and benzodiazepines in depression. We’ll pick up with part II on April 26, where we’ll explore efforts to treat depression and anxiety disorders with the z-hypnotics, particularly eszopiclone.

CHRIS AIKEN: After Ayd’s book, another pioneering psychopharmacologist – Karl Rickels – released a controlled trial comparing the barbiturate phenobarbital, the benzo diazepam (that’s Valium), and the antidepressant amitriptyline in 138 patients with depression. Surprisingly, he found no difference in the outcomes, even when he stratified the patients by anxiety. Studies like this encouraged doctors to continue using gaba-ergic medications in depression, even though they were not classified as antidepressants.

But the definition of depression was also shifting during these times, and Frank Ayd’s book had a lot to do with that. Before Dr. Ayd’s text, depression was only diagnosed in the severe melancholic cases, and these patients were usually treated in the hospital. The patients we now call depressed in outpatient practice were more likely called anxious or neurotic in the 1950’s. The one-percenters were treated with psychoanalysis, but most Americans – about 1 in 10 of them – and that’s 1 in 10 Americans not 1 in 10 people with anxiety disorders – were treated with the gaba-ergic barbiturates, or a barbiturate like compound that became the first psych med to reach blockbuster status: Meprobamate, which was marketed under the deceptively safe, middle-class name: Miltown. It was the success of Miltown that prompted the search for a safer gaba-ergic, which came out in 1960 as chlordiazepoxide (or Librium), the first benzodiazepine.

Turning Anxiety into Depression

Dr. Ayd believed that many of these Americans who were taking sedatives from their PCP or psychiatrist really suffered from depression, and he wrote this book to help them recognize and treat it. It’s rare for a single book to change medical practice, but Ayd’s did, with a little help from a friend. Merk pharmaceuticals had just released the tricyclic antidepressant amitriptyline (Elavil) in the same year Ayd’s book came out, and Ayd had served as a consultant as they developed the drug. Merk bought up 50,000 copies of Ayd’s book and delivered it to every primary care doctor in the country. The investment paid off, and doctors soon had a new explanation for the stressed out, glass-half-empty, worried and demoralized group of patients that have come to our aid for thousands of years: They were depressed.

The strategy paid off for Merk, as amitriptyline went on to become the best-selling tricyclics, even though it wasn’t the first tricyclic to be launched. But it didn’t quite put an end to the gaba-ergic line of treatment. Benzodiazepine gradually replaced barbiturates, and by the 1970’s they were as common as the barbiturates were in the 1950’s: 1 in 10 Americans were taking them. The benzos were faster acting and better tolerated than the antidepressants, so many doctors preferred them for milder cases of depression. Some psychiatrists argued that the benzos should be reclassified as antidepressants, while others countered that they didn’t treat core symptoms of depression and any therapeutic effect they brought about was likely due to the diagnostic muddling of depression and anxiety that was happening all over the place and is still going on.

One of those nay-sayers was a young Alan Schatzberg, who is better known today as the author of two APA textbooks of psychopharmacology. In 1978 Dr. Schatzberg published a paper in the Archives of General Psychiatry – known today as JAMA Psych – where he reviewed 20 controlled trials in a meta-analytic fashion. This was before meta-analyses became popular, but Schatzberg developed his own method, classifying each paper based on whether the benzos came out better, equal, or worse than antidepressants. In the end, the benzos were more often the losers than the winners, and when they did win it was often because the antidepressant was dosed too low. Schatzberg concluded that benzodiazepines helped sleep and anxiety in depressed patients, and that when they did improve mood it was usually a secondary effect of their anxiolytic and sedative benefits. He concluded the benzos did not target core symptoms of depression like anhedonia, low motivation, and psychomotor retardation.

In a most fields that would have killed the argument, but in psychiatry ideas have a way of reinventing themselves, like the antipsychotic-antidepressant combo pills that were popular in the 1970’s and have resurfaced again today. Three years after Dr. Schatzberg’s article a new kind of benzodiazepine was released – the high-potency triazolobenzodiazepine alprazolam, Xanax. This benzos has a triazole ring added to its benzodiazepine structure, which some believe makes them resemble a tricyclic antidepressant, both drugs have a benzene rings fused to other rings.

And so a new series of studies was launched, this time testing alprazolam in depression.

Alprazolam (Xanax) and Depression

KELLIE NEWSOME: Alprazolam – Xanax – has accumulated enough controlled trials in depression that the uninitiated might assume it was an antidepressant. It was used as monotherapy in 21 randomized controlled trials involving 2,700 subjects. These studies compared alprazolam to placebo and/or another antidepressant – usually a tricyclic. Alprazolam generally worked as well as the antidepressant, not just for anxiety but sometimes for the core symptoms of depression as well. The typical dose was 3 mg a day. Not surprisingly, alprazolam worked faster than the antidepressants and was better tolerated.

Alprazolam’s antidepressant potential was given the green light in two Cochrane reviews – which is pretty hard achieve – the Cochrane group is so conservative they usually conclude that there’s not enough evidence to support anything we do. And even Dr. Schatzberg has come around. In the latest edition of his Manual of Psychopharmacology he mentions that alprazolam – alone among the benzodiazepines – has direct antidepressant effects. Alprazolam is not FDA approved for depression, but the FDA allowed the labeling to state that alprazolam is effective for anxiety associated with depression. On the other hand, the FDA recognized that its antidepressant action can go too far. Alprazolam is the benzo that is most associated with manic induction in case reports, and these are mentioned in the package insert as well.

Although the bulk of this benzo-depression research supports alprazolam, some psychiatrists believe that its antidepressant effects are not unique and extend to all benzos. We spoke with Giovanni Fava who recently published a systematic review and meta-analysis of 38 controlled trials of benzos in depression, and he did not think that the antidepressant effects were unique to alprazolam. In his analysis, the research supported other benzos as well, such as the first benzo: Chlordiazepoxide (Librium). Dr. Fava pulled together international team of 6 psychiatrists, including Karl Rickels. They sliced up the data in various ways to try to parse out which kind of depression responded best to the benzodiazepines, but no type rose to the top. Anxious depression was just as likely to respond to a benzo as a tricylic.

However, Dr. Fava did note that most of the depressions in the benzo literature were mild cases, and that limitation is in line with a problem that other authors have noted. Most have concluded that the benzos – including alprazolam – do not work as well in the classic, melancholic depression where patients have early morning awakening, guilty rumination, low appetite, psychomotor changes, and an unreactive mood that is distinct from ordinary sadness. Melancholic depression has always been thought of as the more biologically based of the depressive syndromes, and in some ways the lack of response to a benzodiazepine validates the idea that this really is a unique syndrome that requires a very specific treatment.

Benzos in Depression: The Drawbacks

But the bigger drawback to all this benzodiazepine research is that we have no long-term studies on benzos in depression, and there are clear concerns about dependence, tolerance, and withdrawal – not to mention its abuse liability. Some psychiatrists have argued that the SSRIs are just as bad, with a withdrawal syndrome of their own and suggestive – but not conclusive – evidence of tolerance. But let’s be clear: The withdrawal syndrome on benzos is far worse and more dangerous, and SSRIs do not have a street value.

Concerns about their long term effects and abuse potential probably explains why so few psychiatrists use alprazolam as monotherapy in depression. But many do use benzos when starting an antidepressant. A popular strategy is to start a benzo while waiting for the antidepressant to kick in, and then taper the benzo off after a month or two. There are 10 controlled trials evaluating this strategy, and it does speed up the antidepressant effect, but after a month this kind of augmentation doesn’t seem to make a difference.

The hard part is getting patients off the benzo. You really have to prepare them for that step because these drugs do have rewarding effects. About 1 in 10 patients with depression receive a benzo with their initial antidepressant prescription, and about 1 in 8 of those are still on the benzo a year later. But while this strategy does have risks, it probably helps patients remain in treatment and tolerate any initial side effects on the antidepressant – many of which, like nausea, insomnia, and restlessness – improve with benzodiazepines. If you’re going to do this, alprazolam is probably the way to go as it has the most evidence, but if your concern is addiction you may want to use a benzo that’s slower to act and less rewarding like lorazepam or oxazepam.

CHRIS AIKEN: But before we wrap this up – let’s pause and remember what many of us were taught in medical school: Depression is a known side effect of benzodiazepines, particularly with long-term use. It’s also a problem with barbiturates. Depression on barbiturates is one of the reasons that these drugs have fallen out of favor in epilepsy.

Besides the fact that these medications are sedating and slow the mind and body down, benzodiazepine have psychological effects that are not desirable. They blunt emotions, although you could argue that SSRIs do the same thing. But here’s a difference: SSRIs tend to improve the ability to accurately recognize emotional expression in other people’s faces, while benzodiazepines do not. We know this from studies of normal people, as well as depressed people on SSRIs, and last year we covered a new study that showed benzodiazepines impair the ability to recognize angry faces. That might sound like a pleasant la la land at first, but imagine if every time you made someone angry you were unable to recognize what was happening – your friends and lovers might dismiss you as out of touch or get so angry at your non-responsiveness that they explode.

Benzodiazepines also impair cognition with long-term use, and I’m particularly concerned by the finding that they impair problem solving ability, as poor problem-solving skills are one of the main reasons patients take them. They seem to decrease learning during psychotherapy, and one wonders if they decrease the ability to learn from challenging stressors and process difficult emotions across the board.

In part II of this podcast we’ll get into the latest iteration of the gaba-depression story, as investigators have looked at whether the z-hypnotics improve non-sleep symptoms in patients with depression, anxiety, schizophrenia, and pain disorders. You’ll recognize a lot of the same themes in that line of research – questions about whether the benefits are direct effects of the medication, or indirect effects of improving sleep. Whether the benefits are short lived, or long term, and conflicting evidence about whether the z-hypnotics treat depression or cause it.

For myself, I’m in an uncomfortable position with all this. We’ve said a lot of good things about the benzos today, and we’ve pointed out a lot of problems. All of our drugs do some good and do some harm, but we can do a lot of harm when we’re lured into thinking that a certain psychiatric medication is as wholesome as apple pie – which, when you think about it, really isn’t that wholesome.

KELLIE NEWSOME: It is an uncomfortable position, but there’s still a lot you can do with this knowledge, however uncertain it is. Like here’s a case I saw this week. A 55 year old man with panic disorder came to me on 6 mg a day of alprazolam. That seems high, but it’s actually within the normal dose range for panic. Anyway, he was doing really well but had read that benzos were bad for the brain and wanted to come off of it. So we slowly tapered down the alprazolam over the past year. His panic did not return, and he’s pretty free of anxiety, but something else happened – he gradually got more and more depressed as we lowered the dose. We could try an antidepressant at this point, but he never tolerated them very well. So we raised his alprazolam back up a bit and his depression got better.

Look out for that kind of reaction in your own practice. And while we don’t recommend long-term benzodiazepines as a routine therapy for depression, there are always refractory cases that don’t respond to anything else.

CHRIS AIKEN: We are also frankly undecided on whether alprazolam is the one to go with when a patient has depression and you need to use a benzo – say – to treat comorbid panic disorder. 21 studies is pretty solid support for alprazolam’s antidepressant effects, but there are also reasons to avoid this benzo – alprazolam has a higher abuse potential than other benzos, a higher overdose potential when taken with opioids, and it’s long duration and hepatic metabolism make it less desirable in the elderly and medically ill where short acting benzo that’s not metabolized through the liver like lorazepam or oxazepam is a better choice.

KELLIE NEWSOME: And now, for the word of the day….the N-Back Test

The N-Back Test

CHRIS AIKEN: Have you ever struggled with a dense treatise on psychopharmacology and wished you could increase your IQ by 3 points? That’s what the N-Back Test promises to do. Now, if you don’t think 3 points doesn’t sounds like a lot, take it from Steve Martin – who has a self-declared IQ of 142. Here’s what the comedian describes in his book Pure Drivel:

[Steve Martin Quote]

The N-Back test is a test of working memory developed in the 1950’s, but it gained attention in 2008 when a controlled study in the Proceedings of the National Academy of Sciences found that regular practice with this test increased fluid IQ – which measures problem-solving and reasoning skills. The original paper was criticized for methodological weaknesses, but it was followed up by 20 controlled studies that together suggest that N-back practice has a small but real effect on intelligence, equivalent to a 3-4 point elevation of IQ.

The test is a little like the card matching game “Memory” or “Concentration” that children play, but instead of remembering cards you have to remember letters. Here’s how it works. Someone reads or shows you a series of letters, and every time they read one that is a repeat from three letters ago you press a button. Let’s try it out now….


CHRIS AIKEN: To get it right, I have to keep the last three letters in my head while listening to the new letters, looking for patterns, and dumping the old letters from my working memory as new ones come in.

The difficulty can be adjusted by requiring you to recognize items from 2 times back, 3 times back, 4 times back, etc… hence the name N-Back.

There’s also a dual-N back test, where you are presented with letters and shapes at the same time, and you have to recognize when either the letter or the shape is a repeat of one that was shown 2 times ago. It’s like playing 2 chess games at the same time. Most of the IQ research was actually done with this version. I didn’t find the game very pleasant, so have decided to stick with my natural born IQ.

The N-back test is important because the cognitive gains it brings, although small, are generalizable. In other words, even though it is just a test of working memory, regular practice brings about improvements in broader cognitive skills. The N-back is also commonly used to test for working memory in psychiatric disorders, and next week’s podcast will feature a new app that lets you measure and track cognitive functioning in patients with 4 well-validated tests: The choice reaction time task, the digit symbol substitution test, the trail making test, and the n-back test. If you want to try it out, google N-Back. There’s a lot of free versions on line.


Today is the first Monday of the month, and you know what that means – a new episode of Pocket Psychiatrist is out which you can share with your patients. On today’s episode, Dr. Greg Sazima teaches a mindfulness breathing skill. Follow us on twitter, where Dr. Aiken is continuing his marathon of posting one practice-changing study a day. Today’s study looks at whether an isomeric metabolite of sertraline/Zolfot called dasotraline treats ADHD.

If you’d like to earn CMEs for these Carlat articles, stop by our website and use the promo code PODCAST to get $30 off your first year’s subscription.

The Carlat Report has operated free of advertising since 2003. Well, OK, we did include two ads in this episode but they were for historical purposes only and we assure you that money did not change hands. One was a 1950’s public service announcement for Atarax called The Relaxed Housewife, and the other was an educational video from the 1970’s encouraging cardiologists to prescribe Valium to help their post-MI patients engage in rehabilitation and exercise.