Sunday, November 12, 2017

More on Benzodiazepines (Like Xanax).

The topic of benzodiazepines will just not go away.  At the top and bottom of this post - I include a number of book covers from my library on the topic from the last 30 years.  The publication dates are 1983, 1985, and 1990.  I wrote a brief review for the Psychiatric Times and included my unedited version  on this blog from earlier this year.  My overall message was that benzodiazepines as a group are great for very specific indications.  In fact for some indication like detoxification from alcohol or sedative hypnotic drugs and catatonia they are life saving.  The majority of benzodiazepines are not prescribed for those indications.  They are prescribed primarily as add on medications for anxiety and insomnia.  Their use is limited by tolerance and addictive properties that are expected from a medication that reinforces its own use.  It is also problematic to prescribe them to any population of patients where alcohol use is prevalent and not expect significant drug interactions or abuse.

I really wanted to include a graphic from the paper listed below (1) from JAMA Psychiatry on the prescribing rates of benzodiazepines in patients being treated for depression.  I will post it if I get permission.  That graph illustrates the growth in benzodiazepine prescribing from 2001 to 2104.  During that time the fraction of patients taking a benzodiazepine in addition  to an antidepressant rose from 6% to 12.5% of the depression treated patients.  Subgroups were analyzed and psychiatry came across as a the top prescriber of benzodiazepines across all years.  Other practitioners and specialists came in under the curve prescribed by psychiatrists. 

 A recent anxiety disorder diagnosis was a strong predictor of concurrent use of benzodiazepine use with 24.1% of patients with a recent unspecified anxiety disorder diagnosis and 39.1% or patients with a panic disorder diagnosis taking both the benzodiazepine and antidepressant.  At 6 months 45.6% of the antidepressant + benzodiazepine and 48.1% of the antidepressant monotherapy were still taking an antidepressant.  After initial adjustments 12.3% of the simultaneous benzodiazepine and antidepressant users received long term benzodiazepines with 5.7% taking them for one year.  The prescribed benzodiazepines were almost all high potency including alprazolam (43.9%), lorazepam (26.3%), and clonazepam (21.8%).  About a tenth (13.5%) of the patients got a limited supply of for only 1-7 days).

The authors generally conclude that benzodiazepine prescribing seems to be consistent with current guidelines.  These suggest that a Cochrane report that concomitant benzodiazepine use with antidepressants increased the short term antidepressant response and decreased the drop out rate attributable to antidepressant side effects.  I would think that would be offset at least as much by guidelines suggesting limited use.

The overall strength of this report is that it is a study of a very large insurance database population of 684,100 new antidepressant users and 81,020 simultaneous antidepressant and benzodiazepine users. They give a breakdown of antidepressant classes (overwhelmingly SSRIs).  Only 12.7% and 16.5% (respectively) of the patients in each class were treated by psychiatrists.  Interestingly 77.4% and 80% of each class had not received any form of psychotherapy, although it is conceivable that at least some patients were being seen by therapists outside of the database.  As noted psychiatrists were more likely to prescribe combination therapy.  The authors speculate that may be due to referral patterns and psychiatrists seeing patients with more severe anxiety and depression, training patterns in psychiatry, or more familiarity with benzodiazepine pharmacology.  I think a more likely factor is chronicity and the fact that psychiatrists tend to see more patients with chronic anxiety and temperamental forms of anxiety that are not taken into account in DSM-5 nosology.

Despite the large N, there are many drawbacks to database studies like this one.  I think it is useful in terms of the basic pharmacoepidemiology of prescriptions but it doesn't say anything about symptoms severity and many of the practical issues involved with benzodiazepine prescribing (like substance used disorders) are eliminated by the study protocol.  The authors do a good job of describing the downsides (use disorders, falls/fractures, motor vehicle accidents) of benzodiazepines in their discussion.  I would have included cognitive problems and tolerance. It also does not address the optimal way to address combined anxiety and depressive disorders. My biggest concern is the current "evidence based" fad of diagnosing anxiety and depressive disorders using symptom rating scales like the PHQ-9 (Patient Health Questionnaire-9) and the GAD-7 (Generalized Anxiety Disorder 7-item).  In a primary care setting they pass for a diagnosis.  In a psychiatric setting they short circuit any analysis of the etiological factors of anxiety or depression.  Instead these disorders are conceptualized as disorders that that require a basic medical treatment and they resolve.  That is a gross oversimplification.     

But the main limitation should be evident - we do not have a specific enough diagnostic system with reliable objective markers.  In that context a large N doesn't mean as much unless we know how many subtypes there are and the associated treatment parameters.

George Dawson, MD, DFAPA


1:  Bushnell GA, Stürmer T, Gaynes BN, Pate V, Miller M. Simultaneous Antidepressant and Benzodiazepine New Use and Subsequent Long-term Benzodiazepine Use in Adults With Depression, United States, 2001-2014. JAMA Psychiatry. 2017;74(7):747–755. doi:10.1001/jamapsychiatry.2017.1273


  1. Right now the opiates are getting the press, but, once a couple of celebrities or other well known citizens succumb to ODs with Xanax or Klonopin as primary substances of abuse, then, we'll see the spotlight shine on the Benzos.

    Oh, and read in the front pages of the recent Current Psychiatry that it won't be long before the stimulants join in the mix.

    Amazing how my colleagues are a few years behind my experiences from what I have been seeing in community mental health clinics these past few years to now. It is the Trifecta: especially in middle aged women, getting started on opiates for some vague pain issue, then adding the benzos like Xanax to "treat" what is likely subclinical opiate withdrawal mistaken as "anxiety attacks", and then craving stimulants to wake them up from the opiate/benzo fog each morning.

    Makes John Belushi sigh in envy...

    Joel Hassman, MD
    Board Cert Psych MD

    1. This type of polypharmacy is very common - there are often multiple prescribers. Some very common scenarios that I encounter:

      1. Benzodiazepines(BDZ) to cover opioid withdrawal.
      2. BDZ to augment the euphorigenic properties of opioids prescribed for pain.
      3. BDZ used to get high on methadone prescribed for methadone maintenance.
      4. BDZ and z-drugs to treat insomnia from heavy alcohol use (the 3AM awakening).
      5. BDZ to treat AM alcohol withdrawal symptoms (prevent shakes, sweats at work).
      6. Stimulants to be able to stay up and drink more alcohol.
      7. The classic speedball combination of cocaine or meth with heroin (or any combination of what is considered the prescription equivalent).
      8. Stimulants to balance the excessive sedation from benzodiazepines +/- opioids happens but it is less common. I do see people taking stimulants for excessive sedation from atypicals + BDZ + mood stabilizers for a "bipolar disorder" diagnosis.

      Lots of polypharmacy out there and the combinations are risky.

      My apologies for the late post. Tech problems.

  2. Another problem with the study is that I didn't see any information about dosing. There is a world of difference between someone taking say 2 mg of clonazepam tid and someone taking 0.25 mg qHS. Without having the dosing information the inferences that can be drawn are rather limited.

    1. Agree - the authors use a very unusual methodology that looks at whether the dose is at, above, or below the modal dose but no amounts are specified. Actual mg amounts would have been preferable. I would also like to have seen an example of their calculation. I assume they just look for the most frequent dose.

      Sorry for the delay in posting - tech problems.