Monday, April 29, 2019

Deprescribing - Same Job With A New Spin

During my tenure as an acute care psychiatrist, I had to reconcile a lot of medications. I was doing medication reconciliation before the term was invented for the electronic health record (EHR). The process basically involves trying to figure out what medications the patient was really taking before they were admitted to the hospital. It could be very easy if there were no preadmission medications. On the other hand it could be extremely complicated. There were days when I had to sort through two or three shopping bags full of medications, talk with the patient’s pharmacist, talk with several specialists who were prescribing medications, and talk with the patient’s primary care physician. Even after that long process, I often estimated initial dosages based on the patient's recollection of what they had been taking and how much. I also had to make fairly rapid decisions about whether or not large numbers of medications may have been more harmful to the patient than helpful. Some patients had lists of medications containing 10 to 20 unique medications.

Sometime in the past 10 years the concept of deprescribing medication came up. It is fairly unique term as indicated by the bar graphs below that are drawn based on the references per year to the term. It started out the geriatric literature because elderly people are more sensitive to lower doses of medications and polypharmacy relative to younger and healthier populations. There is actually a list of medications called Beer’s list, that highlights medications that may be more problematic in older adults. It is the intellectual property of the American Geriatrics Society and I can’t reproduce it here. It basically contains classes medications that are known to be problematic in older adults such as anticholinergics and sedative hypnotics. Consistent with that concept - the geriatrics literature has focused on rational pharmacology and the need to reduce the medication burden in some cases the specific pharmacodynamic burden of prescribed medications.  Goal of this post is to look at some of the techniques I typically use to identify polypharmacy - related problems and respond.

In determining whether deprescribing should occur or not I think it is useful to look at hierarchy and I have outlined the following points:

1. In the case of the patient on polypharmacy who is tolerating multiple medications well and they appear to be effective strongly consider doing nothing:

Being an expert in psychopharmacology - doing all the reading and listening to the experts often doesn’t translate into the real world setting very well. There’s no better example than the patient on multiple medications who frequently has a history of numerous or prolonged hospitalizations and who appears to be taking “too many medications”. They could be multiple medications from the same class or different classes. It is easy to take a look at that list of medications and imagine how they came about but with our current fragmented medical record system it would only be an imagining.  It is too high of a risk to stop polypharmacy just based on general principles if the patient is doing well. I am familiar with many cases where changes were made and the patient became markedly destabilized and ended up back in long-term hospitalization. These the cases that never come to light in the literature were populations rather than outliers are studied.

2. Acute medication side effects: 

In the case of acute side effects changes need to be made based on the urgency involved. Worst-case scenarios would include serotonin syndrome or neuroleptic malignant syndrome where the serotonergic or dopaminergic medications need to be stopped abruptly. That would not occur in typical clinical scenarios but in the emergency setting it is necessary. What clinicians typically face is multiple medications from the same class. When that original guideline was made back in the 1990s classes were a lot more general than they are now. For example, in those days antidepressants were general class instead of SSRIs, SNRIs, and others.  These days combination antidepressant therapies are relatively common and research articles can be found that look at the addition of bupropion to a standard antidepressant or mirtazapine to a standard antidepressant. Beyond that trazodone might be added to those two antidepressants bring the total to three. This can occur commonly in clinical practice and also can be a source of the patient noted in number 1 above.

Numerous side effects can result from polypharmacy like sedation, headaches, nausea, and cognitive problems that probably indicate the total amount medication needs to be decreased or at least one of the medications could be stopped. The medication I frequently encounter that is prescribed at very high doses resulting in sedation is Venlafaxine ER.  There are areas of the country where very high doses of this medication are prescribed in excess of 350 mg per day (225 mg per day) is considered the FDA recommended max dose. Almost uniformly these patients improve with less venlafaxine and there is less confusion about medication side effects versus depression.

3. Chronic medication side effects:  

Some of the most serious long-term medication side effects include weight gain, metabolic changes including metabolic syndrome, diabetes mellitus, nephrogenic diabetes insipidus, hyperlipidemia, and movement disorders. In many cases the medications being used that lead to the side effects have been the only ones that will that work and even gradual changes may result in destabilization the patient. Some of these transitions between atypical antipsychotics or atypical antipsychotics and mood stabilizers result in a significant medication burden and risk for increasing side effects. It is critical that the transition is actually made to the new set medications.

4. Rare but serious medication side effects: 

Looking at both neuroleptic malignant syndrome and serotonin syndrome, the literature frequently states that these acute life-threatening disorders occur around times of medication transitions. Trying to keep the load on both serotonergic and dopaminergic systems low during these transitions is one of my goals but I can’t really find any scientific literature to back it up. Literature out there tends to be case reports and that includes literature suggesting that medication transitions are associated with the acute disorders.

5. Interrupted medication transitions: 

I frequently see people who are on full doses of two and often three antidepressants. When I take their history there was a plan to add the new antidepressant and then taper and discontinue the old one but for some reason the old medication was not stopped.  This often happens in the outpatient setting and many times it is due to the patient not knowing that the old medication should be stopped or not getting a specific schedule to taper and discontinue it.

6. Polypharmacy: 

Polypharmacy can be highly problematic. It happens in just about every class of psychiatric medications. As an example, Adderall XR is designed to produce a concentration curve that is equivalent to Adderall immediate release dosed twice a day and yet I commonly see people taking Adderall XR either more than once a day or combined with an afternoon dose of Adderall immediate release. There are similar combinations of antidepressants, antipsychotics, mood stabilizers, and benzodiazepines. In a controlled setting where I practice I can make the necessary medication changes and follow-up the patient frequently. If that occurs in the outpatient setting there needs to be a plan in place for frequent follow-ups as well as active collaboration with the patient and the family.

7. Pharmacokinetic problems: 

The most common pharmacokinetic problem I encounter is people who abruptly stop Lamotrigine and resume the full dose.  Since lamotrigine began its psychiatric applications I have been in touch with the manufacturer many times and was advised that if the patient stops the medications for more than three or four days, the standard titration of lamotrigine needs to occur. It is fairly common for me to hear from people that they go off lamotrigine for a week or two and then resume the full 200 or 400 mg dose. I often see them after they have been on that resumed dose for one week.

The prototypical pharmacokinetic polypharmacy problem was SSRIs that were CYP2D6 inhibitors combined with tricyclic antidepressants (CYP2D6 substrates). The original reports of severe arrhythmias in some cases death from tricyclic antidepressant toxicity was the initial impetus for psychiatric interest in pharmacokinetics and drug interactions. I still see people today who are getting amitriptyline or nortriptyline in combination with fluoxetine or paroxetine and there has been no clear concern about those potential interactions.

8. New medical problems that impact prescription patterns:  

Acute renal and hepatic problems can directly impact the patient’s drug metabolism and dosing requirements or ability to take a specific drug.. One of the best examples I can think of is a case of 40 year old man who was taking gabapentin for anxiety and chronic pain. He was seen by an internist and started on a statin for dyslipidemia. Four days later when I saw the patient he was delirious and completely disoriented. He also had the significant ataxia and sedation. He was evaluated immediately and blood tests showed that he had acute renal failure that was believed to be secondary to the statin. The statin and the gabapentin were discontinued and within days he was back to his baseline.  If he had been on any other medications with primary renal clearance those would have been discontinued at same time.

9.  Correcting the medical side of things:

If the psychiatric medications are being taken incorrectly, there is a good chance  that the polypharmacy for heart disease, hypertension, diabetes mellitus, and asthma/COPD are also being taken incorrectly if they have been taken at all. It is problematic when a person has a disabling mental illness and they are left to take several doses of medication at different times of the day by themselves. When I started out in psychiatry, I could make a public health nursing referral at any time by sending in a form to the appropriate agency.  The next day, and RN would be at the patient's apartment setting up their medications, taking their blood pressure and pulse, and assisting them with managing their medications for the psychiatric disorder as well as all of their chronic medical problems.  That service ended with the rationing of all services to people with severe psychiatric disorders, making it much more likely that these medical conditions will not be as stable as they should be when they see their psychiatrists.  The is both a problem for the patient and the psychiatrist but also an opportunity to correct things.  

These are a few examples of the hierarchy of problems that occur with polypharmacy and in some cases standard pharmacy and how they can be approached. There apparently some groups out there at this time were trying to establish a hierarchy of how medications can be discontinued and when they should be discontinued. Like most cases in medicine in the extreme it is obvious but anything less than that is more difficult and it takes a lot of time to figure out. One thing that might be useful would be to consider drug combinations that are commonly prescribed as a baseline and look for polypharmacy being defined as anything beyond that.

One thing is for sure - the old rule about never prescribing two drugs from the same class - no longer applies.

George Dawson, MD, DFAPA

Sunday, April 14, 2019

Kratom - Don't Believe the Hype

The CDC came out with a brief 2 page report on kratom deaths 3 days ago (1). That was all it took for the Twitterati to proclaim that there were many more deaths from alcohol and I suppose there was a post about even more deaths from cigarette smoking but thankfully I missed that one. When I pointed out that it was clearly an addictive drug and lifelong disability (a very significant problem) may be the issue - the defenders of kratom stepped up and talked about how harmless it is and also how it is advantageous for people who cannot afford medication assisted treatment (MAT) (buprenorphine preparations, methadone, or naltrexone extended release injections) for opioid use disorder (OUD). The expected personal attacks and sarcasm followed.

Kratom is an interesting compound because like many psychoactive botanicals there is a history (2). Kratom itself is basically leaf material from the kratom tree (Mitragyna speciosa). The leaves can be smoked or chewed. They can also be dried and powdered. The powdered form is what is typically available for sale. The powder can be packaged in capsules and taken orally, brewed into a tea, or rendered into a syrup and formed into pills. Fresh leaves can be chewed with or without betel nuts. Kratom has been used in Malaysia since the 19th century to “heal opium addiction”. A recent paper referenced a study of kratom users that were using an estimated 4-8 g/day (8). Converting based on typical leaf content means that these users would be exposed to a maximum of 120-180 mg mitragynine and 1.1 - 3.4 mg 7-hydroxymitragynine.   Rӓtsch suggests in his text that “in studies with mice, even extreme dosages of 920 mg/kg did not produce any toxic effects”. He describes “self experiments” in the literature suggesting that kratom can be both stimulating like cocaine and sedating like opium. The only comment on addiction is “The alleged kratom addiction is a Thai cultural phenomenon” (p. 367).  Like most intoxicants in the modern era there is progression to intravenous use.  Although that is currently rare, there are case reports of intravenous use of kratom extracts.

The CDC document describes a series of deaths in 11 states between July 2016 and June 2017 and an additional 27 states from July to December 2017. The data set was from the SUDORS (State Unintentional Drug Overdose Reporting System) and consisted of 27,338 overdose deaths, 152 (0.56%) of which were kratom positive. There is no standard postmortem toxicology protocol and as previously noted that is problematic in determining the drugs present in these analyses. As shown by the table from this report in 91 cases kratom was considered the cause of death, but numerous other substances were present. In seven cases kratom was the only substance noted in postmortem toxicology, but additional substances cannot be ruled out.

A report in the New England Journal of Medicine, looked at 15 cases of death (4) associated with kratom in Colorado. In this series the authors used more rigorous toxicological analysis with high-performance liquid chromatography – mass spectrometry. Whole blood mitragynine concentrations were noted between 16-117 ng/ml and up to 4800 ng/ml. In this series, 14 of 15 deaths had multiple drugs leading the authors to conclude that these deaths were kratom related.  This series of cases illustrates the importance of toxicological analysis and specifically plasma levels of the drug to correlate with various toxidromes and post mortem toxicology.

The leaves of Mitragyna speciosa, contain multiple alkaloids including mitragynine, 7-hydroxymitragynine, paynantheine, speciocilatine, and speciogynine. The crude alkaloid extract consists primarily of 66% mitragynine and 2% 7-hydroxymitragynine. The extraction process may be protective against toxicity for many people that brew the leaves into a tea, chew the leaves, or ingest the powdered leaves as capsules but even then the concentration of these alkaloids may vary from species to species. Counting on an inefficient extraction process for safety is probably not the best idea.  The other property of the raw material is that the alkaloids are mixtures of  opioid receptor agonists and antagonists that may determine the net effect. Searching the way these products are sold there is really not much about concentration of any associated alkaloids other than mitragynine.  The plant itself contains more than 40 unique alkaloids (8).

Until recently, the pharmacology of mitragynine and 7-hydroxymitragynine were unknown. There is research to suggest (5) that opioid receptors mediated the primary effects. Both compounds had binding affinity for the mu opioid receptor (MOR). They were also active in tissue essays and blocked by naloxone.  Some of these effects were inconsistent between laboratory species. Activity was reported at a number of non-opioid receptors as well. The pharmacology of mitragynine and 7-hydroxymitragynine is now well-characterized. Recent studies show that mitragynine is a partial agonist at the human mu opioid receptor (hMOR), and a competitive antagonist at the human kappa opioid receptor (hKOR), and an antagonist at the human delta opioid receptor (hMOR) but with very low potency. The authors studied these compounds against all three human opioid receptors looking at both functional activity (EC50 and IC50) and binding affinities (Ki) and discovered they were consistent across those experiments. They concluded that mitragynine (0.233 μM) And 7-hydroxymitragynine (0.047 μM) had significant binding affinity for hMOR. The remainder of the paper focuses on medicinal chemistry theory, specifically how opioid -like compounds that bias intracellular signaling toward G proteins rather than β-arrestin may be better candidates for opioid analgesics with low addiction potential and better side effect profiles and possibly antidepressant activity. They synthesize a number of analogues and look at their agonist activity at hMOR. The authors conclude that the psychoactive activity of Mitragyna is most likely due to their action at hMOR. They also point out that due to the competitive nature of the alkaloids the gross effects will be due to that balance of agonism and antagonism.

The alkaloid and methanol crude extracts of kratom are both inhibitors of CYP3A4 and CYP2D6 in vitro. No specific components have been identified with this activity and there has been in vivo confirmation (8).

Another paper (6) looks at the “unanticipated toxicity” of kratom. This group looked at the LD50 of mitragynine, 7-hydroxymitragynine, and heroin. The LD50 is a measure of acute toxicity and what single dose will kill half of the research animals. In this case mice were used and the researchers were surprised to find that an intravenous dose of either mitragynine or 7-hydroxymitragynine were as lethal as heroin. No lethal doses were observed for oral dosing in a range of 6.25-50 mg/kg. the lethal intravenous dose was midpoint in that range. Researchers observed that the mice appeared to die from respiratory depression within 10 minutes of direct exposure. In the surviving mice many were noted to have seizures in the first 20 minutes.  In a separate review, the authors point out that with a typical 8 g dose of kratom powder, the levels of 7-hydroxymitragynine, may be too low to cause a pharmacologically relevant effect at the opioid receptor. 

The research on kratom has elucidated receptor activity in opioid receptors. The activity is complex but the mu opioid receptor is clearly involved and is the likely site of the psychoactive effects and the application of opioid substitution in people with addictions. The receptor effect is complicated and likely involves more than the mu opioid receptor. The research also suggests that activity at murine and human opioid receptors are not equivalent. Persons acquiring kratom in the powder form need to consider that the ratio of mitragynine to 7-hydroxymitragynine likely varies with species and source. The 7-hydroxymitragynine is 52 times as potent as mitragynine at the MOR and 13 times as potent as morphine. Products made from extractions will be more potent.

All of this information should create skepticism in prospective kratom users. As addiction psychiatrist I can attest to the fact that it is addicting and with any addiction there is a tendency to escalate the dose. Many people with addictions as noted in the above table are using multiple substances some of which are also agonists at the opioid receptor. If you are considering kratom as a treatment for opioid addiction or chronic pain there are much, much safer and effective ways to proceed.

George Dawson, MD, DFAPA


1. Olsen EO, O’Donnell J, Mattson CL, Schier JG, Wilson N. Notes from the Field: Unintentional Drug Overdose Deaths with Kratom Detected — 27 States, July 2016–December 2017. MMWR Morb Mortal Wkly Rep 2019;68:326–327. DOI:

2: Rӓtsch C. The Encyclopedia of Psychoactive Plants: Ethnopharmacology and Its Applications. Park Street Press. Rochester, Vermont, 2005. pages 366-367.

3: Olsen EO, O’Donnell J, Mattson CL, Schier JG, Wilson N. Notes from the Field: Unintentional Drug Overdose Deaths with Kratom Detected — 27 States, July 2016–December 2017. MMWR Morb Mortal Wkly Rep 2019;68:326–327. DOI:

4: Gershman K, Timm K, Frank M, Lampi L, Melamed J, Gerona R, Monte AA. Deaths in Colorado Attributed to Kratom. N Engl J Med. 2019 Jan 3;380(1):97-98. doi: 10.1056/NEJMc1811055. PubMed PMID: 30601742.

5: Kruegel AC, Gassaway MM, Kapoor A, Váradi A, Majumdar S, Filizola M, Javitch JA, Sames D. Synthetic and Receptor Signaling Explorations of the Mitragyna Alkaloids: Mitragynine as an Atypical Molecular Framework for Opioid Receptor Modulators. J Am Chem Soc. 2016 Jun 1;138(21):6754-64. doi: 10.1021/jacs.6b00360. Epub 2016 May 18. PubMed PMID: 27192616; PubMed Central PMCID: PMC5189718.

6: Smith LC, Lin L, Hwang CS, Zhou B, Kubitz DM, Wang H, Janda KD. Lateral Flow Assessment and Unanticipated Toxicity of Kratom. Chem Res Toxicol. 2018 Nov 16. doi: 10.1021/acs.chemrestox.8b00218. [Epub ahead of print] PubMed PMID: 30380840.

8: Kruegel AC, Grundmann O. The medicinal chemistry and neuropharmacology of kratom: A preliminary discussion of a promising medicinal plant and analysis of its potential for abuse. Neuropharmacology. 2018 May 15;134(Pt A):108-120. doi: 10.1016/j.neuropharm.2017.08.026. Epub 2017 Aug 19. Review. PubMed PMID: 28830758.

9: Post S, Spiller HA, Chounthirath T, Smith GA. Kratom exposures reported toUnited States poison control centers: 2011-2017. Clin Toxicol (Phila). 2019 Feb 20:1-8. doi: 10.1080/15563650.2019.1569236. [Epub ahead of print] PubMed PMID:30786220.


FDA Guidance On Lead and Nickel Exposure from kratom products:  Link

"Based on these test results, the typical long-term kratom user could potentially develop heavy metal poisoning, which could include nervous system or kidney damage, anemia, high blood pressure, and/or increased risk of certain cancers."

Graphics Credit:

1.  Mitragynine and 7-hydroxymitragynine were done with ChemDoodle.

2.  Table is from the CDC per reference 3 and public domain.

Sunday, April 7, 2019

More On Conscious States and Suicide

“Did you remember all of that noise I was making in the bathroom?  I was trying to kill myself.”

The person I was talking with had been discharged from a hospital about two months ago.  He was admitted there because of an exacerbation of a mood disorder and possible psychosis. The main reason he was admitted from the emergency department was suicidal ideation. That is the most frequent indication for hospital admissions in the United States. Even then who does and does not get admitted is controversial. It is common for persons to be sent to the emergency department by their families or outlying facilities where there are legitimate concerns only have the patient deny the problem and get released from the hospital. There is a lot of drama involved because one of the decision points is whether or not suicidal person needs to be placed on legal hold and treated on an involuntary basis. This frequently leads to speculation about the true nature of what a person says or alternatively accepting "no suicidal thinking" at face value and dismissing them. 

I think it also highlights the significant limitations of interviewing people and adequately    understanding their conscious state. The best example is the rating scale approach which is really somebody’s idea of what the optimal interview questions might be to assess a suicidal person. The commonest depression checklist is the PHQ-9 (1).  Item 9 in the PHQ 9 involves suicidal thinking and the rating is as follows:

Thoughts that you would be better off dead, or of hurting yourself
0 – not at all

1 - several days
2 - more than half the days
3 - nearly every day

Depending on where you practice clinics have different conventions about this item and how it needs to be approached. Any elevation usually leads to a more intensive assessment of suicide potential. That typically involves a clinical interview but also could involve the use of another checklist. It should be apparent that this item is a focused on the approximate frequency of suicidal thinking. It assumes that the patient can actually report this and that it is more significant than other metrics like the intensity of thinking. For example, is one extremely intense thought about suicide more significant and potentially lethal than thinking about it frequently but easily dismissing those thoughts? This is one of the basic limitations of any assessment of the person’s mental status. Clinical interviews and rating scales are very crude approximations of a person’s conscious state. Assessing someone’s potential for suicide is a clear example. There is also the notion of rating scales being “quantitative” measures and they are not. There is an entire field of research suggesting that these “measurements” lead to greater precision and I doubt that is true.

All of that brings me back to the first patient. Here he is somewhat annoyed that nobody seemed to realize on an inpatient psychiatric unit that he was trying to kill himself. At the same time he made every effort to conceal that fact while he was hospitalized. He only disclosed it months later after his mood and associated cognitive processes had stabilized. It reminds me that I also have talked with many people who were intent on killing themselves and presented themselves as being very well so that they could be discharged and attempt suicide. The popular literature is full of stories about people who reassured their families or appeared to be doing well only to carry out a planned suicide attempt. This is clearly a high risk conscious state that can escape detection and lead to very high risk attempt or death.

“The gun just went off.”

I talked to many survivors of gunshot wounds that were self-inflicted. In large trauma hospitals, psychiatrists are consulted by surgery services who have successfully treated the patient. The psychiatrists job is to assess the patient and determine whether or not they need further acute psychiatric care or they can be discharged home. I generally ask for a very detailed description of what happened including the type of firearm used, the time of day, the associated thought process, the overall psychiatric context, and the sequence of events just before the firearm goes off. The common explanation that I have heard is a recollection that someone was pointing a loaded gun at themselves and that at some point it "just went off". There is no recollection of a conscious effort to pull the trigger. Numerous secondary analyses are possible including that it is just a rationalization against self-harm or an attempt to avert psychiatric hospitalization. In keeping with the theme of this post - there is also a possibility that the patient’s conscious state at the time of the suicide attempt was so chaotic that it cannot be recalled or reconstructed. There is precedent for that state and that is delusional depression. If the patient is clearly delusional all of the usual deterrents like fear of dying, intense dislike of pain, not wanting to harm the family, and religious beliefs no longer apply. The standard risk analysis for suicidal thinking no longer applies. There is a delusional process with associated emotions that lead to very high suicide risk.

“I felt real bad about what happened 50 years ago and so I stabbed myself.”

The delusional process can be very subtle. Psychiatrists are typically taught to pay attention to hallucinations and classic forms of delusions. Those types of psychotic thinking are fairly obvious. In the case of depression and some forms of psychosis the delusion can be very subtle. An example might be feeling guilty about a trivial event from a long time ago. Everyone can relate to that kind of guilt or embarrassment but what if it is suddenly linked to the idea that death is preferred to the emotional burden of that trivial event. People in their 50s, 60s, and 70s could focus on events that happened when they were in middle school or high school that might start to disrupt their lives and lead to suicidal thinking. In the example given a severe suicide attempt occurred by self-inflicted stab wound over a trivial incident happening in the eighth grade. The patient was unable to recognize that this was a delusional thought process until the depression and psychosis had been adequately treated.

These examples all highlight how a person can go from being no risk at all for suicidal behavior to being at very high risk. The changes are subtle and they might not be apparent to the person experiencing them. The risk analysis models that are used are all linear and additive and do not capture the conscious states of people who become suicidal. The limited consciousness theories that we currently have would suggest that it is really not possible to experience the conscious state of another person in the transition to high suicide risk is probably a good example.  Even the best possible definition of empathy fails if the person cannot recognize the state that the psychiatrist is trying to reflect back to them. 

Time domain is another perspective on the fluidity of conscious states both in the case of suicidal thinking and substance use disorders. It is common for a person to describe themselves as becoming a person that they never wanted to be associated with both substance use disorders and suicidal thinking.  They are able to see those patterns in retrospect but not at the times they occur.   

It may be apparent that suicidal thinking can be a transition from a questionable belief to certainty. I listed a few of these beliefs in a previous post. A common one is “people would be better off without me”. In the early stages most people can examine that thought and conclude that it is at least partially false based on their relationships to the people in question and the assessment of their realistic value to those people. With time and continued emotional intensity any objective assessment of their value in relationships might diminish and disappear. At that point they are in a very high-risk state because they believe in the statement that “people would be better off without me”. Clinicians are often taught to ask about deterrence to suicidal ideation, but they are rarely taught to assess the degree of belief a person has in high-risk suicidal thinking.  There are non known ways to determine is a person who is delusional or quasi-delusional about suicidal thoughts is disclosing those thoughts or hiding them.

What can clinicians and patient do in these circumstances?  My previous posts suggests that an analysis of the thought patterns can be useful. I routinely review those ideas with people I see who have suicidal thoughts.  At some point the goal would be to see if talking about suicidal thoughts in this way would improve the level of resistance to these thoughts and make it less likely that people will act on them. I also believe that a public health message should discuss the same approach,  So far the only public health measure seems to be advice on calling suicide hotlines or crisis lines. 

I have had several people who I know as friends let me know that they have been able to analyze these thoughts on their own and come up solutions to contain these thoughts and get enough emotional distance from them to the point that they were no longer bothersome.  I know it can be done and encourage public health officials to take it to the next step.   

In closing, this post emphasizes a unique conscious state or states associated with suicidal ideation and suicide attempts. Nothing in this post should be construed as interview or treatment suggestions.  A more comprehensive understanding of  suicidal thinking and behavior requires more than a rating scale approach or risk factor analysis.

George Dawson, MD, DFAPA


1.  PHQ-9 is copyrighted by Pfizer, Inc. Full rating scale is visible at many sites by searching on PHQ-9.

Monday, April 1, 2019

Vigilance Is Required For Adequate Informed Consent

One of the main reasons for me writing this blog is to discuss medication safety and adverse effects. This is no small task since the orientation of most research and clinical work is making a diagnosis and where appropriate selecting the correct medication. A lot of work goes into those determinations but an equal amount of work needs to be directed toward managing common adverse effects and preventing serious adverse effects/events of the medication. In order to do that practitioners need to realize that they are treating a genetically diverse population. That means there will always be a subgroup of people who cannot tolerate medications. There will also be a subgroup of people who respond very well to medications and variations in between.

A good example is selective serotonin reuptake inhibitors (SSRIs). These medications are commonly first-line agents for depression and anxiety. About one person out of every seven will not be able to tolerate them. The same is true of every major class and psychiatric medications. That is a source of frustration for practitioners because it means that even if the correct diagnosis is made the optimal medication may not be available because the person cannot tolerate it. In the case of SSRIs, a person may not be able to tolerate any other medication in that class.

In addition to genetic heterogeneity affecting the pharmacokinetics and pharmacodynamics of drugs, patients are also taking other classes medications and have other medical morbidities that may contraindicate the use of certain psychotropic medications. In the case of bupropion, any history of eating disorder or current eating disorder, any seizure history, and any traumatic brain injury that may predispose to seizures are contraindications to using that medication. Appropriate medical care requires that practitioners are aware of all of the contraindications and precautions listed in FDA package inserts for medications.

What about the case of a healthy young person with no medical comorbidity who needs treatment for depression or anxiety. What is important for that person to know in order to safely take that medication? What should be explicitly discussed with them? What constitutes adequate informed consent? What can be done in the time a practitioner has to make the diagnosis and educate the patient? These are not trivial questions since most practices are scheduling patients every 15 minutes in some cases new patients are seen for 30 minutes. All that makes it seem like psychopharmacology is a very easy job but the information transfer during the sessions is critical and there is plenty of evidence that it is inadequate.

There is an excellent discussion of the problem by Rajnish Mago, MD in the references (2).  He considers a number of ways to assess and manage adverse effects including open-ended questions, checklists administered to the patient, structured interviews by clinicians, and spontaneous reports by patients. His review of this literature shows that the sensitivity and specificity of all of these methods are lacking and some of them take so much time they could not be applied clinically. He gave as an example of structured interviews administered by clinicians that take up to 60 minutes just to determine the adverse effects.  The other problem with determining the adverse effects is whether or not they can be attributed to the medication that has been prescribed. In clinical trials, researchers often estimate whether adverse effects/events are due to the medication or not. In practice that is a difficult determination due to both placebo and nocebo effects.

Vital signs and laboratory measures can also constitute adverse effect measurements. In psychiatry liver function tests, renal function including plasma creatinine and estimated GFR, CBC and ANC, TSH, T4, basic metabolic profile, electrocardiograms, and EEGs are all metrics that can be followed to determine medication effects. Vital signs should also be routinely done on anyone taking medications to check for effects on blood pressure, heart rate, and heart rhythm. Those measurements constitute more specific measures that could lead to clinical action.

There has been some emphasis on measurement-based practice that involves the extensive use of rating scales.  Many of the authors in this area seem to mistake rating scales as both objective and quantitative measures when they are neither.  I remain unconvinced that this is the best approach and think that it has resulted in the over medication of large numbers of people who have a certain rating scale score but no diagnosis.  There has been very little discussion of the analysis of longitudinal data and what these rating scale scores actually mean. The controversy in this area applies to both the diagnostic indices as well as the side effect indices. My approach in dealing with side effects is that they should be completely eliminated wherever possible and that irreversible adverse effects need to be avoided.

Dr. Mago’s paper reviews recommendations for improving adverse effect reporting and clinical trials and also his recommendations for clinical practice that basically come down to telling patients about all adverse effects that occur at a rate 5% greater than placebo and all potentially serious for life-threatening adverse effects even if they are rare. I am in complete agreement with those recommendations, and even have an approach that can work. It does require a degree of vigilance on the part of the clinician.

The attached table suggests why vigilance is required.  Psychiatrists need to be more vigilant than most physicians because no patient is expected to get complications or die as a direct result of our treatments. That historically has led the field to have a lower threshold for monitoring for potential side effects.  Given all of these constraints and the complexity of the situation is there a way to provide adequate informed consent about the medications that we prescribe?  I am talking about all physicians here and not just psychiatrists.  I think there is and I will walk through my process.

1.  I preface my remarks with my experience prescribing the medication: For example with common medications like naltrexone I will give an estimate of the percentage of people that tolerate the medication very well and the percentage that stop taking it and why. That sets an  expectation that a medication may or may not be well tolerated as a probability statement and that some people stop it because of adverse effects.

2.  I encourage people to do their own research. Anyone can pull up the FDA package insert on any medication these days by Googling: "[medication name] FDA package insert".  I tell them what to expect. I also tell them that I should be able to explain anything they find on the Internet about the medication if they have any questions or concerns. In the case of polypharmacy scenarios, I point out that the extremes of low doses and high doses of two agents and all of the combinations in between can be found this days with stated results ranging from very positive to very negative for the entire range of doses.  I also have the position that they can take as long as they want to do their own research and that in the meantime - no medication needs to be prescribed.

3.  I will discontinue the medication at any time and I am very explicit about that.  I do not expect anyone to "get used to" side effects because in my experience too many people get used to side effects and live with them on a long term basis.  I provide examples to illustrate the idea of living with side effects.

4.  I give them the MedlinePlus handout on the medication or show them how to get it.  The MedlinePlus handouts are really a product of American Society of Health-System Pharmacists (ASHP).  I have emails both into MedlinePlus and AHSP inquiring about the process they use to determine the bullet points for the listed precautions and tiers of side effects. At the time of this post I have not been contacted by either organization.

5.  I tell them the common side effects, contraindications to the medication, and precautions - specially if they are on a medication or have a medical condition that is flagged in the precaution section.

A good example is the antidepressant duloxetine. The contraindications are straightforward - uncontrolled angle closure glaucoma and recent use of a monoamine oxidase inhibitor (MAOI). Beyond that clinical worsening and suicide risk, hepatotoxicity, orthostatic hypotension and syncope,  serotonin syndrome or neuroleptic malignant syndrome, abnormal bleeding, activation of mania/hypomania,  seizures, effect on blood pressure,  clinically important drug interactions, hyponatremia,  use in patients with concomitant illness, urinary hesitation and retention, and  laboratory test abnormalities are all listed as warnings and precautions.  The most important one that I typically discuss is liver function abnormalities due to alcohol use.  Although this is not a contraindication - the package insert basically says that a person with this problem should not take duloxetine. I have had people want to take the medication despite this warning and have had to discuss a monitoring plan with them as well as letting them know that my read of the package insert is that this plan carries with it more risk than one that adheres to the recommendations by the FDA.   The package insert also lists two different groups of Treatment Emergent Adverse Reactions in tables.  Both are more common than placebo and occur in 5% or 2% of patients in clinical trials.  There is a subsequent list of side effects defined with these frequencies:

"Reactions are categorized by body system according to the following definitions: frequent adverse reactions are those occurring in at least 1/100 patients; infrequent adverse reactions are those occurring in 1/100 to 1/1000 patients; rare reactions are those occurring in fewer than 1/1000 patients." 

It takes a lot of practice to get the above informed consent discussions down to 10 minutes, especially when polypharmacy and other specific problems need to be addressed at the same time. Even then, it is important to present the information every time. Shortcuts can result in a patient experiencing an uncommon side effect and that leads to self doubt on the part of the physician of the form:

"If I had mentioned that problem like I usually do would it have led to a more timely intervention to reduce the adverse effect?"

That is a question that most physicians don't want to keep asking themselves and it is why my vigilance is high.

George Dawson, MD, DFAPA


1: Bloom R, Amber KT. Identifying the incidence of rash, Stevens-Johnson syndrome
and toxic epidermal necrolysis in patients taking lamotrigine: a systematic
review of 122 randomized controlled trials. An Bras Dermatol. 2017
Jan-Feb;92(1):139-141. doi: 10.1590/abd1806-4841.20175070. PubMed PMID: 28225977;
PubMed Central PMCID: PMC5312199.

2: Mago R. Adverse Effects of Psychotropic Medications: A Call to Action.Psychiatr Clin North Am. 2016 Sep;39(3):361-73. doi: 10.1016/j.psc.2016.04.005. Review. PubMed PMID: 27514294.

3: Tse L, Barr AM, Scarapicchia V, Vila-Rodriguez F. Neuroleptic Malignant Syndrome: A Review from a Clinically Oriented Perspective. Curr Neuropharmacol. 2015;13(3):395-406. Review. PubMed PMID: 26411967; PubMed Central PMCID: PMC4812801.