Silexan, an extract from Lavender with pharmaceutical properties, is one of only two treatments with a large effect size in generalized anxiety disorder. We rank all the treatments by strength and look to see if any of them can work when the SSRIs do not.
Published On: 8/10/2020
Duration: 26 minutes, 2 seconds
How well do psychiatric medications work in generalized anxiety disorder? And what do we do when they don’t work? 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.
KELLIE: I’m holding in my hand a paper from 2016. It’s a metaanalysis of treatment resistant anxiety disorders by Beth Patterson, and I hoped it would have some answers for this problem that I see everyday in practice. But the conclusion is that nothing works. Nothing.
Anxiety is paradox. Patients feel it physically, so they naturally expect that a physical treatment like a medication will help. But it’s in the anxiety disorders that our medications have some of the smallest effect sizes. In this month’s Carlat report we featured an article on a German medication called Silexan that has beaten the odds in a few large head-to-head trials generalized anxiety disorder. Silexan worked much better than paroxetine, equally well as lorazepam, and consistently better than placebo. That’s all exciting but what good is a medication that’s only available by prescription in Germany and a dozen or so other European countries? Well we were able to find a source for this medicine and in the online edition you can learn how to order it and use it.
AIKEN: Silexan is a proprietary extract of Lavender oil made by Schwabe pharmaceuticals. Schwabe is a German company that has been making over the counter remedies since 1866. What they do is identify natural sources of medicinal products, like ginkgo or in the case of Silexan the lavender plant whose oils and aromas have been used for anxiety for centuries. They then cultivate specific strains of the plan to intensify the therapeutic ingredients, and extract them in a proprietary way.
You may also know them if you follow the tardive dyskinesia literature. They make tebonin – a ginkgo extract with the generic name EGb 761 that has evidence in tardive dyskinesia. Tebonin came out in 1965 and has mainly been used for dementia and age related cognitive decline. It has neuroprotective effects and antioxidant effects that suggested it may treat tardive dyskinesia, and was tested in 3 randomized controlled trials totaling 300 patients where it worked with a very small p-value: 4 zeros. By comparison our FDA approved medications were tested in about 400 patients each, and arrived at a p-value that is 100-times smaller than tebonin’s: only 2 zeros.
KELLIE: Lauging. “I’m sorry but you’re reminding me of a joke. I was at a medical conference once and this guy had a t-shirt on that read “My p value is smaller than your p value”
CHRIS: Well it’s interesting that tebonin has the smaller p value here, but is often mentioned as an afterthought – if at all – in clinical papers on tardive dyskinesia. That’s not true in neurology by the way – the American Academy of Neurology included tebonin in its list of 4 standard treatments for tardive dyskinesia.
KELLIE: Why do you think that is – that psychiatry gives natural products a short shrift?
AIKEN: I think in psychiatry we’re just more careful. The public is very skeptical and suspicious of what we’re doing – giving medications for the mind – and we do have a history of going too far in the wrong direction like with frontal lobotomies. And in America we are particularly skeptical of natural products because they’ve tended to grow like infestations in our capitalist economy – in fact it’s because of these snake oil remedies that the FDA rose to power as a regulatory agency – first in the 1930’s to regulate safety and then in the 1960’s to regulate efficacy.
KELLIE: But the FDA doesn’t regulate natural products much at all. As long as they don’t make libelious claims or cause major harms, they can do whatever they want. But in Europe – and in particular in Germany – it’s different. There they respect natural products and regulate their safety and efficacy. Many are available by prescription only, and 70% of general practitioners in Germany prescribe herbal remedies. Herbal medicine is a required part of the medical curriculum in Germany. In Germany natural medicines like Silexan and Tebonin are under much tighter regulatory control. In Germany, herbal products have to prove their quality, safety, and effectiveness before entering the market – it’s also this way in the UK, the United Arab Emerits, and the Kingdom of Bahrain. In America, any herbal product can enter the market, and they can only be taken off the market if the FDA proves that they’re not safe.
So in America when we think of natural products we think of scandals where poorly manufactured herbal concoctions that were barely effective to begin with caused big problems, like tryptophan which people sometimes take for anxiety and depression. Well in 1989 the FDA recalled a tryptophan product from Japan that contained contaminants that were causing eosinophilia-myalgia. But fast-forward 30 years and the FDA is doing the same thing with medications that it approved: Both Ranitidine and Metformin were recently recalled for contamination with a carcinogen, NDMA.
AIKEN: Much of that has to do with the trend toward overseas manufacturing of drugs in the past 10 years. It’s much harder for the FDA to regulate that. For example, the FDA inspects American factory through surprise visits, but when they go to inspect overseas factories they have to authorize the trip through the local governments so their visits are announced weeks in advance and the companies have time to clean up any problems.
KELLIE: OK so bottom line: this German preparation worked with a large effect size in generalized anxiety disorder. Depending on the study, you report that that effect ranged from medium – 0.5 to large – 0.9. Can you put those numbers in context – what do they mean?
AIKEN: An effect size is how much better than placebo the medicine worked. They range from small, 0.2-0.5. That’s what we see with SSRIs in depression, around 0.3-0.5 depending on whether we look at all studies or just the published ones. When we include all studies we’re including some flawed ones, like they might have enrolled people with only mild symptoms who weren’t really depressed and do just as well on placebo as the drug, so that’s why you see such a small range. SSRIs also land around 0.3 for generalized anxiety disorder. In this small range, you’d barely notice the effect with your own eyes.
Next is a medium effect size. That’s 0.5-0.8. This is where the casual observer can see a difference. Benzos for generalized anxiety fall here at 0.5. When all psychiatric treatments are piled together – including medications and psychotherapy – the average effect size as 0.5, which is about the same as the average effect size in all of medicine as a whole. Medium.
KELLIE: So what’s a large effect size?
AIKEN: Not many treatments in psychiatry have a large effect size. There’s stimulants for ADHD, ketamine for depression; those studies nearly always average out in the large range. Now, depending on the study we might also include benzos for acute panic or antipsychotics for acute psychosis, and now Silexan for Generalized Anxiety Disorder.
KELLIE: Let’s finish up with all the treatments for Generalized Anxiety. This all comes from meta-analyses so you can’t really compare these effect sizes head to head. For example if one study enrolled patients with severe illness, or one study had a particularly weak placebo because they weren’t giving much supportive care to their subjects, then you may see very different effect sizes for the same drug. But those caveats aside here’s where the numbers fall:
We’ll start with the lowest effect – vortioxetine Trintellix. Although it’s promoted by Stephen Stahl and others as a good antidepressant for anxiety, it’s effect size in Generalized Anxiety didn’t even cross into the small range. It’s 0.12, negligible
- Buspirone 0.17… negligible
- Vilazodone Viibryd 0.26, Small
- SSRIs 0.33
- SNRIS 0.36
- Hydoxyzine / Vistaril 0.45
- Benzos 0.4-0.5
- Psychotherapy 0.5
- Silexan 0.86
- Quetiapine 1.6
We pulled those numbers from half a dozen meta-analyses, several of which were done at Duke University. For the most part all of them arrived at the same numbers, but where there was a discrepancy we went with the more recent paper, for example pregabalin came in at 0.5 in 2007 but 0.38 in 2017 when more trials were added in.
KELLIE: I noticed the SNRIs like venlafaxine and duloxetine were more effective than the SSRIs.
AIKEN: Yes 0.33 vs 0.36, and that’s a trend we tend to see in depression as well the SNRIs ring in with an effect size that’s just slightly above that with SSRIs, but not enough to make a meaningful difference in practice.
KELLIE: And Vistaril – hydroxyzine – worked better than I’d expect
AIKEN: Hydroxyzine is an antihistamine that’s been out since the 1950’s. It’s actually FDA approved for “relief of anxiety and tension associated with psychoneurosis,” which I suppose is an antiquated term for generalized anxiety disorder. It’s often viewed as ineffective by clinicians, who seem to reserve it for patients with addictive tendencies who request as needed meds, but maybe we need to reconsider that. The main limitation in that effect size is that it was only based on short term studies. The main reasons to avoid hydroxyzine are sedation, dry mouth, qtc prolongation, and anticholinergic effects, and while those anticholinergic effects might raise concerns about memory problems it was tested against lorazepam and was much more favorable in terms of cognition and balance than the benzo.
KELLIE: But what really stands out to me there is that quetiapine – Seroquel - is the most effective. Why is it not FDA approved in generalized anxiety then?
AIKEN: The company tried to get FDA approval in 2009 but the FDA rejected it – not because it wasn’t effective – it clearly is – Quetiapine’s effect is based on 3 studies with 1,533 patients, and the margin of error around that 1.6 effect size is pretty tight – 1.6 is the average, and the real effect is somewhere between 1.0 and 2.2, so either way it’s large.
KELLIE: OK so why didn’t they approve it?
AIKEN: Safety. The FDA decided that generalized anxiety disorder was not a severe enough condition to warrant risking it with a medication that can cause diabetes and tardive dyskinesia among other things. And this gets back to the origins of Generalized Anxiety Disorder in 1980. If you’re interested scroll back to our November 2019 series on anxiety. Generalized Anxiety Disorder was originally conceived of in 1980 as a mild condition where psychotherapy was appropriate and medications were questionable. Then the SSRIs came out in 1987, and with their greater tolerability they gained approval and widespread popularity for generalized anxiety disorder.
KELLIE: So when would you use quetiapine for anxiety?
AIKEN: Only if it was very severe and disabling. Generalized anxiety occurs on a spectrum, from everyday worry to disabling misery. And the FDA’s concern is well justified, because the pharmaceutical industry likes to push the envelope of that spectrum. If quetiapine had been approved they would have been pitching it for everyday anxiety to general practioners everywhere. You might also consider quetiapine for patients that have a lot of anxiety and another disorder where quetiapine is indicated – like bipolar disorder.
KELLIE: It’s interesting that the FDA views depression as severe enough to warrant approval with quetiapine, but not anxiety.
CHRIS: That may reflect culture more than science. When the world health organization surveyed psychiatric disability in 2018, they found that 15% of people with psychiatric disorders were disabled. The odds of having a disability were about the same for Depression (3.5) as generalized anxiety (3.1). A lot of PR has gone into convincing people that depression is a serious illness; anxiety, not as much.
KELLIE: OK but this podcast is on treatment resistant anxiety disorder. And that was my frustration – I found nothing that worked! Does Silexan work in treatment resistant cases?
AIKEN: Not that we know of. It has a large effect, but that’s in people just starting treatment. Treatment resistant anxiety is defined much like treatment resistant depression – failure to have a meaningful response to 1-2 therapies, usually antidepressants. And here’s where things get interesting, although not very inspiring. The paper you’re holding looked at whether adding something to an SSRI to could work when anxiety doesn’t respond to SSRIs. And like you said they concluded that nothing worked. And most of those studies involved which medication?
KELLIE: Let me see… quetiapine, quetiapine, olanzapine…. Quetiapine. IN fact of the dozen studies they looked at over half of them involved atypical antipsychotics, and they didn’t work. AIKEN: Surprising given that large effect size! We can’t always explain these things, other than treatment-resistant anxiety disorders must be very hard to treat. Sometiems meta-analyses are so conservative in their rigor though that they miss important signals, like in this paper they disqualified studies if the patients were taking too many other psychotropics or didn’t publish exactly the kind of metric they were looking for. I get it those can be weaknesses, but I’d still like to see the results. So we pulled up the individual papers to see if any of them showed efficacy on their own terms. And one did. It was a study of pindolol augmentation in panic disorder – this paper by the way looked at generalized anxiety, social anxiety, and panic disorder.
KELLIE but before you go into that result – I’m skeptical. Why did they exclude it from the metaanalysis?
AIKEN: It wasn’t for good reasons in my opinion. First, they nixed it because the authors didn’t list a primary outcome. They used 6 rating scales, and the problem there is by chance they could get a positive result on just one of those rating scales and claim success. But the study was actually positive on all but 1 of those six rating scales: it was negative on the Hamilton depression scale, and since it’s a study of anxiety I doubt they would have made that the primary one. The other reason they excluded it was that the authors didn’t specify whether there drop outs were in the treatment or placebo group. And yes, a study could seriously misrepresent their results if the medication looked really positive in the end but almost all of the drop outs were taking the medication. Well, in this study they only had one patient drop out, and he dropped out because he moved abroad, so this technically really doesn’t apply.
KELLIE: It reminds me of what the US Immigration Services has been doing recently. They deny applications for immigration if they have any blank field on their application – even if it doesn’t apply – like if they leave their middle name blank because they don’t have one, or leave the names of their siblings blank because they are an only child…
AIKEN: OK but we’re trying not to get political on this podcast
KELLIE: Right. I’m sorry. Back to pindolol. Did it work?
AIKEN: Yes. Pindolol 2.5mg tid successfully augmented fluoxetine in panic disorder. The study was small, only 25 patients. But it’s a all we got – nothing else that was tested in treatment resistant anxiety worked – not pregabalin/Lyrica, quetiapine or other atypical like olanzapine, ziprasodone, risperidone; not clonazepam/Klonopin
KELLIE: Wait a minute, you’re telling me that adding a benzo doesn’t do anything in anxiety disorders?
AIKEN: This was a large study in social anxiety. They took patients who did not respond to sertraline and tested whether augmenting with placebo, klonopin, or switching to venlafaxine would help – and your instincts are right kellie the konopin did help, but it wasn’t a big enough effect to be statistically significant.
KELLIE: That’s not what I see in practice. I mean patients generally report a big difference when we add in a benzo.
CHRIS: Yes but this paper is just reporting on an anxiety rating scale. Benzos may be doing all sorts of things – including their known effects on the reward pathways – that aren’t measured on an anxiety rating scale. But it’s questionable whether their rewarding effect is a beneficial outcome. It certainly improves adherence though.
On the other hand, eszopiclone/lunesta 3mg at night did improve anxiety symptoms – not just sleep – in a large controlled trial where it was added to escitalopram for generalized anxiety. That study was in 2008, and it’s important because it suggests that anxiety improves when sleep gets better – which is certainly true in my own life.
KELLIE: Why wasn’t that one in the meta-analysis?
CHRIS: They didn’t have treatment resistant anxiety – on the other hand they were taking escitalopram 10mg for 3 months and most hadn’t fully recovered on it.
KELLIE: So let’s get to the bottom line. Pindolol is FDA approved in hypertension, but it has CNS effects – it’s a serotonin 1A 1B antagonist. Would you use it in treatment resistant anxiety?
CHRIS: Yes for panic disorder, but I wouldn’t rush into it. I might consider an MAOI first – those have evidence in panic, and in social anxiety they have a large effect size than the SSRIs. For treatment resistant Generalized Anxiety I’d certainly explore psychotherapy and this new Silexan. But I have to acknowledge that Silexan – despite it’s large effect size – has never been studied in treatment resistant anxiety. But other than pindolol nothing else has worked, and when you scan through the treatments in this paper – antipsychotics, anticonvulsants, benzos… pindolol is probably the safest of the bunch. The problem is that pindolol’s study is small and not replicated, but it’s all we got, and while it was just in panic disorder pindolol also small trials in treatment resistant depression.
KELLIE: And now for the word of the day Alienist.
An Alienist is a 19th century term for a psychiatrist, and the word in Latin means “outsider or foreigner, and the word came into the English language during the 19th century from the latin by way of the frenchwhere the adjective aliene (insane) gave rise to the noun alieneste – a Dr who treats the insane. It was also the title of a popular novel of a 19th century psychiatrist in NYC. It’s a curios term and to our modern years it would seem that people back then stigmatized mental illness viewing it as though they were aliens from out of space. That is not the case ,the use of the word derives from the fact that they viewed people with mental illness as alienated from themselves and from society.
To get the full wrap on how to use Silexan, check out our August issue online.
Join us next week for an interview with Jonathan Silver on Traumatic Brain Injury.
Got questions or ideas? We’d love to hear from you: firstname.lastname@example.org.
And your subscription helps us stay free of industry influence so we can keep bringing you unbiased information you can trust.
Got feedback? Take the podcast survey.