A Realistic Comparison of SSRI and Heroin WIthdrawal

With all the quotes in the media about these withdrawal syndromes – I thought I would add a few facts.  These facts are based on science, my 22 years in acute care psychiatry, and another decade at a large facility specializing in substance use disorders.  I have been involved in the care of thousands of people with these disorders, and the care has improved substantially over that period of time largely due to the availability of Medications for Opioid Use Disorder (MOUD).  Before that there was a major problem treating anyone with an Opioid Use Disorder (OUD) if they were not enrolled in a methadone maintenance program. 

To cite a few examples, in acute care psychiatry it is common to receive emergency admissions of people with severe depression or psychosis.  Many have associated drug and alcohol problems.  In the case of heroin and other opioids, there was a rule extending up into the early 2000s that unless a person was enrolled in a methadone maintenance program, only acute detox could be done with methadone.  Additional medications were offered like clonidine to cover hyperadrenergic symptoms of opioid withdrawal.  Clonidine was often the only detox medication. The situation was not better in substance use disorder (SUD) treatment programs who often had an array of “comfort medications” designed to treat all of the peripheral symptoms of withdrawal.  The problem with this approach was twofold: first it was rarely adequate to treat acute withdrawal. Second, when the patient was discharged in a week, they had ongoing symptoms of withdrawal that essentially guaranteed an immediate relapse to opioid use.

The introduction of buprenorphine (Suboxone, Subutex, Sublocade) resulted in a marked improvement in the quality of care in this scenario. The acute and chronic withdrawal symptoms of opioids could be adequately treated and the risk of relapse mitigated. During the time of this transition, the large SUD treatment center where I was working went from using buprenorphine for acute detox to buprenorphine maintenance treatment.  I was responsible for treating depression and anxiety in these patients.  I observed that no matter what treatment I prescribed the anxiety, insomnia, irritability, and depression persisted for months as protracted opioid withdrawal symptoms until they were treated with buprenorphine.  At that point the withdrawal symptoms resolved immediately and completely.  The depression, anxiety, insomnia, and craving for opioids all resolved.  That massive improvement in care cannot be emphasized enough.  I went from discharging people after 1-3 months who I knew would relapse immediately to confidently discharging people who were stable and had a much better chance to recover.

I am posting this introduction because the historical window for this innovation was very brief and I doubt that many physicians ever saw it.  The advantage for patients was so clear that the medical director of the treatment program where I worked changed it from an abstinence-based program to MOUD.  The research at the time was clear.  MOUD saved lives by decreasing relapse rates, accidental overdoses, and decreased risk of infections (HIV, HCV).

In my previous post, I examined the rhetoric of comparing selective serotonin reuptake inhibitor (SSRI) withdrawal to heroin withdrawal.  But what about the popular myths and the science?  The popular view of opioid withdrawal is that once the acute phase is over – it is over.  The popular view promoted with SSRI withdrawal is that it is typically universal, severe, and long lasting.  Neither of those views are accurate.

The time course of symptoms is outlined in the diagram at the top of this post.  Acute symptoms have a characteristic pattern and can start as soon as hours after a last dose of heroin.  In the case of antidepressants, the time course is more dependent on the mechanism of action and half-life of the medication being studied. Since opioids are all mu opioid receptor (MOR) agonists – there are no known medications in this class that do not cause withdrawal, but what accounts for the differences in severity and percentages of people affected is not known. Pharmacokinetics and pharmacodynamics likely play a role – I will defer that discussion to later post.

The main qualitative differences in opioid and SSRI withdrawal are the lack of cravings and observable physical symptoms in SSRI withdrawal.  Opioids reinforce their own use and for that reason have street values.  SSRIs do not and have no street value. The withdrawal symptoms from SSRIs are largely subjective but that does not mean they are not real or serious.  

In the case of SSRIs and other antidepressants, the mechanism precipitating withdrawal is thought to be a precipitous fall in extraneuronal serotonin.  All SSRIs are orthosteric inhibitors of the serotonin transport protein (SERT).  Orthosteric inhibitors act like serotonin at the same binding site on SERT.  Allosteric modulators bind at a site that is topographically distinct from the site that the endogenous ligand (in this case serotonin or 5-HT) binds to.  Escitalopram has an additional effect on SERT as an allosteric modulator.  Trazodone, vilazodone, and nefazodone are all allosteric non-competitive inhibitors of SERT (1-4).  Withdrawal reaction have been described by both but not to the same degree as heroin withdrawal.   

In the case of heroin withdrawal, a series of studies done from 1962 to 1969 showed very high relapse rates (>90%) in heroin users who had been treated in a controlled environment. Thise studies led to early experiments with MOUD and eventually methadone maintenance.  A direct comparison of the likelihood of moderate withdrawal symptoms for each category is 3-31% for SSRIs and 100% for opioids.   Those numbers are qualified by several caveats.  First, opioid withdrawal has been studied for a much longer time, is more well characterized, and fewer medications are involved.  Second, study methodology for the SSRI withdrawal is much more varied from looking at short term randomized controlled studies to surveys of long-term use selected for withdrawal symptoms.  The former will underestimate withdrawal effects but identify a drug attributable effect by subtracting out placebo and the latter will overestimate effects .  The overestimate can be compounded by not subtracting a nocebo effect and publicity effects that may increase nocebo.  In clinical practice, using antidepressants with lower withdrawal risk – I found the incidence to be closest to that of Henzler, et al (7) at about one in six or seven people.    

In conclusion, discontinuing many medications can cause a withdrawal syndrome. Not all withdrawal syndromes indicate an addiction and some can be life-threatening.  They are difficult to study for several reasons.  First, the response to discontinuing the medication varies significantly from person to person even at the same dose and duration of use. That range is significant from no effect to severe and in some cases (alcohol, sedative hypnotics) life-threatening effects.  That includes a placebo response where that can be safely implemented.  Second, there is also a lack of high-quality evidence for many of these syndromes. The best evidence is for alcohol and sedative hypnotics because they are obvious and have been treated for decades.  But even then there are consensus guidelines such as the ASAM guideline on benzodiazepine tapering.  Third, there is limited standardization across guidelines.  Fourth, there are pharmaceutical limitations.  In some cases, very small doses of medication are needed to complete the protocol that may require a liquid form, pill cutting, or substitution with an equivalent medication with a longer half-life.  In some cases, all of these changes may not be enough.  Fifth, in clinical trials where medications were tapered and discontinued there may not be enough reported details to replicate the protocol in clinical practice (5).  Sixth, the outcomes of discontinuation of a medication are complex and include resolution of adverse drug effects, relapse to the treated condition, a unique withdrawal syndrome, or a rebound effect involving the physiological systems that were being treated – like rebound tachycardia after stopping beta-blockers.  Seventh, context is important especially with medications and substances that reinforce their own use.  Adherence to any tapering protocol on an outpatient basis is much less likely to happen.  Eighth, the medicolegal dimension may be a concern.  Because of the all of these factors, there is a lot of uncertainty involved in any tapering and discontinuation protocol.  There is seldom a protocol that will work well for everyone. Because of this risk some guidelines advise clinicians to seek legal or administrative consultation when attempting these protocols.  There is additional risk if relapse to the original condition occurs after a medication has been successfully stopped.

Despite all these concerns tapering and discontinuing medications is foundational medicine in any medical specialty. None of these problems are unique to psychiatry or medication for mental disorders.  As interns most physicians learning how to detoxify acute care patients with substance use disorders – the most common condition remains alcohol use disorder.  In their respective specialties – they learn more about how to do this with specific medications used in their specialty. That has resulted in protocols that can differ from hospital to hospital in the same town. In some cases, the protocols differ in the same hospital over a period of years.  To cite one example, I am aware of a hospital that used oxazepam followed by diazepam and then chlordiazepoxide or phenobarbital as their detox agents from alcohol and benzodiazepines. In some of the standard orders, anticonvulsants were also used to minimize seizure risk.

As more societies and government agencies get involved there is a gradual move to standardization.   Even if we get to that point – individual assessments and close monitoring will still need to be done.  The risk of a withdrawal syndrome from any medication is one that is necessary to medically treat various problems.  It is a decision that physicians and patients do not take lightly.  It is important to discuss that potential risk in the informed consent discussion.    

 

George Dawson, MD, DFAPA

 

References:

1:  Plenge P, Yang D, Salomon K, Laursen L, Kalenderoglou IE, Newman AH, Gouaux E, Coleman JA, Loland CJ. The antidepressant drug vilazodone is an allosteric inhibitor of the serotonin transporter. Nat Commun. 2021 Aug 20;12(1):5063. doi: 10.1038/s41467-021-25363-3. PMID: 34417466; PMCID: PMC8379219.

2:  Sanchez C, Reines EH, Montgomery SA. A comparative review of escitalopram, paroxetine, and sertraline: Are they all alike? Int Clin Psychopharmacol. 2014 Jul;29(4):185-96. doi: 10.1097/YIC.0000000000000023. PMID: 24424469; PMCID: PMC4047306.

3:  El-Kasaby A, Boytsov D, Kasture A, Krumpl G, Hummel T, Freissmuth M, Sandtner W. Allosteric Inhibition and Pharmacochaperoning of the Serotonin Transporter by the Antidepressant Drugs Trazodone and Nefazodone. Mol Pharmacol. 2024 Jun 18;106(1):56-70. doi: 10.1124/molpharm.124.000881. PMID: 38769018.

4:  Murray KE, Ressler KJ, Owens MJ. In vivo investigation of escitalopram's allosteric site on the serotonin transporter. Pharmacol Biochem Behav. 2016 Feb;141:50-7. doi: 10.1016/j.pbb.2015.11.010. Epub 2015 Nov 24. PMID: 26621784; PMCID: PMC4724252.

5:  Dirven T, Turner C, Thio SL, Blom J, Muth C, van Driel ML. Room for improvement in reporting of trials discontinuing long-term medication: a systematic review. J Clin Epidemiol. 2020 Mar;119:65-74. doi: 10.1016/j.jclinepi.2019.11.013. Epub 2019 Nov 29. PMID: 31786152.

6:  Goldberg JF, McIntyre RS, Swartz HA, et al. American Society of Clinical Psychopharmacology Task Force on the Deprescribing of Psychotropic Medications. Recommendations for the Deprescribing of Psychotropic Medications: A Consensus Statement From the American Society of Clinical Psychopharmacology Task Force. JAMA Netw Open. 2026 Feb 2;9(2):e260043. doi: 10.1001/jamanetworkopen.2026.0043. PMID: 41739481.

7:  Henssler J, Schmidt Y, Schmidt U, Schwarzer G, Bschor T, Baethge C. Incidence of antidepressant discontinuation symptoms: a systematic review and meta-analysis. Lancet Psychiatry. 2024 Jul;11(7):526-535. doi: 10.1016/S2215-0366(24)00133-0. Epub 2024 Jun 5. Erratum in: Lancet Psychiatry. 2024 Sep;11(9):e11. doi: 10.1016/S2215-0366(24)00253-0. PMID: 38851198.

 

 

 

 

Pills Don't Save Lives - Psychiatrists Do

I am paraphrasing David Healy from a previous post and I am doing it here to emphasize - it's all about the side effects.  Healy's comment serves as a counterpoint to a highly successful multi-decade advertising campaign by pharmaceutical companies.  It began with the first National Depression Screening Day in 1991.  The emphasis  was on identifying and treating depression with antidepressants.  There was no real discussion of antidepressant side effects or the general problem of side effects with most medications.  Since then antidepressant treatment has been conceptualized as comprehensive treatment wrapped up in a pill or capsule.  That bias continues today as various political forces have shifted depression screening from an annual event to primary care clinics.  Some health care organizations and states consider depression screening and serial ratings of depression to be quality markers of health care services despite the fact that there are definite problems with that idea.  Unless there is a highly specific screening test any screening procedure has the potential to expose more people to the side effects of treatment.  There is no highly specific screening test for depression.

A second factor in considering side effects is the physician's role.  Doctors are trained to identify and treat conditions with surgery or medications.  Psychiatrists have additional training in psychotherapy. When you are in your training, the emphasis in on making the correct diagnosis and selecting the medication that will be the most useful.  Even though medical training is long, the longest you might follow any patient might be for a couple of years.  In medical practice you have the ability to see people for decades rather than months or years and how their medical treatment changes over those years.  You also observe first hand the long term toxicity of many medications when you might have only been exposed to that on a theoretical basis during training.  As a practicing physician you are expected to help people deal with the fact that they have side effects and the medications they are using may not be that useful.  In fact, in many cases they may not be beneficial or may be causing more harm than good.

All of that experience with side effects leads clinicians to develop new practices that they were never trained to do.  Very early in my career, I had the experience of treating a person who had been on an antidepressant for about 6 years.  She had headaches and depression and like many people with chronic depression she was in a stressful situation that she could not remove herself from.  She had chronic depression in the context of a chronic stressor that was not going away.  At some point her headaches resolved and her depression improved.  We decided to taper her off the antidepressant.  She came in 2 weeks later and said: "I feel much better.  All of the years that I was taking that medication I didn't realize it, but I felt like I had the flu.  That has now cleared up."  That early experience led me to modify the ways that I discuss medications with people.

I generally tell people that I don't expect anyone to "get used to" a medication.  I often tell them that people may get used to feeling ill rather than develop a tolerance to medication side effects.  I tell them that if they are experiencing any side effects at all to let me know about it and we will decide what to do about it at that time.  I let them know the range of experiences with medications and what they might expect.  As an example, I might say that "60-80% of people might take this medication and not notice that they are taking anything, but 5-10% of people might not tolerate it at all."  I let them know about all of the FDA contraindications, in some cases I review it with them many times.  I discuss the common side effects and usually provide them with the MedlinePlus handout on the medication.  I think it is more comprehensive than most handouts and it gives the FDA black box warnings (in a red box) on the front of every handout.  I talk with them about rare but potentially serious side effects like drug induced liver disease and arrhythmias and what to look for.  In the case of atypical antipsychotics, I discuss movement disorders and metabolic effects.  I demonstrate what the movements of tardive dyskinesia may look like.  I let people know if the medications they are taking are potentially addictive.  I the case of lithium, I let people know about the unique toxicities and the safest possible way they can take it.  In the case of antidepressants, I let people know that they may be difficult to stop due to discontinuation symptoms.

My side effect discussions with people have taught me valuable lessons.  There are people who are placebo responders and nocebo responders.  The nocebo responders develop problems taking any medication, even medications that are generally well tolerated at low doses.  Some of them are aware of the problem and decline any discussion of side effects.  They might say they don't want the MedlinePlus handout because: "If I read about any side effects I will probably get them."  They would rather be surprised.  Whenever I encounter that attitude, I respect their wishes but advise them to contact me if they have any side effects.  I also recall my Forensic Psychiatry lectures during residency.  The instructor advised us that we "could be sued" if our side effect discussions prevented a patient from taking a useful medication and there was an adverse outcome as a result.  I have realized over the years that basing your decisions on whether you could be sued is generally a bad idea because you can be sued for just about anything.  I think that people need to hear about what really happens with psychiatric medications and consider myself to be a good source of information.

I have also found that there is a hearty group of people who decide on their own that they will try to tolerate side effects and not let me know about it despite our discussion.  When I see them in the follow up appointment they will say: "Well you know doc, I had a pretty good headache the first three days on the medications, but I decided to keep taking it to see if it would go away and sure enough on day 4 the headache was gone."  They tell me that even though I advised them to not tolerate side effects and to call me if they had any side effects.  These patients are almost always men with a history of avoiding doctors and not taking care of themselves.  I guess their experience confirms that some people develop a tolerance to side effects but why would you want to?  I was at a large conference on the treatment of anxiety disorders and listened to a renowned psychopharmacologist talk about his technique for treating anxiety disorders with SSRI and SNRI type antidepressants.  His approach was to keep titrating the medication "to the point of toxicity" and then back off to the lower dose.  My experience has taught me that the best approach in non acute situations is to use the lowest possible dose.  That is usually the dose recommended for anxiety disorders and titrate it to the exact point where the symptoms are in remission.  I am never  compelled to increase a medication by a multiple based on the pill size or a drug level based on the aggregate experience of a cohort of people in a drug trial.

I obsess about the hypothetical.  Physicians in practice are aware of trends in the medications that are prescribed and psychiatry is no exception.  Drug interactions have been an area of focus in psychiatry since it was first learned that fluoxetine could inhibit the hepatic metabolism of tricyclic antidepressants and that could lead to antidepressant toxicity.  I treat people who are often on a mind boggling combination of medications for their chronic illnesses and psychiatric disorders.  I routinely run those lists through one or more computerized drug interaction software packages.  The software is inconsistent and I often have to look up the case report or study that suggest a specific interaction or problem.  I have to make the decision to accept or reject what the software is telling me.  The QTc interval or the interval on the electrocardiogram that corresponds with the total time of ventricular contraction and relaxation has been a major concern since the approval of ziprasidone.  It has been complicated lately by the FDA concern that citalopram may prolong the QTc interval in some people to a significant extent.  I screen people with electrocardiograms if it appears that their clinical status or total medication burden may lead to prolongation of the QTc interval.          

In some cases a concern for the hypothetical requires some inductive reasoning.  Current textbooks, literature, and standard prescribing references create the illusion at times that everything is known about a medication, it is just a matter of finding it.  There are plenty of examples where that is not true or where there is a lot of uncertainty about when a medication can be safely and effectively prescribed.  To illustrate, consider a hypothetical situation of patient with bipolar disorder who may benefit from taking lithium.  For a time during my residency training the renal toxicity of lithium was openly debated.  Nephrologists at the time certainly believed it was nephrotoxic but there were large series of patients who were described with minimal signs of renal toxicity.  Clinical practice treating patients with severe bipolar disorder has lead me into situations where I have treated patients on, during and after dialysis and kidney transplantation.  The estimation of glomerular filtration rate (GFR) by 24 hour urine collections was also problematic.  That has been greatly improved by the practice of using calculated GFRs.   I have no doubt at all that taking lithium for a period of time can lead to renal failure in a portion of patients taking it.  Anyone prescribing lithium needs to be aware of this fact and take all measures necessary to minimize episodes of lithium toxicity and exposure to other nephrotoxins.  In some cases like NSAIDs, the toxins are well known.  In other cases like tenofovir, the interactions are not known and in fact you can scan an entire FDA approved package insert and might find no references to lithium.   Making that decision may take hours or a weekend of study to figure out the best course of action.

I hope that I have made the case for psychiatric medications needing a careful analysis of side effects before they can be initiated and continued.  The decision to take medications is a serious one.  In 29 years of practice I have not met a single person who told me that they liked to take medications.  The decision to take medications often comes down to having tried everything else and realizing that a major change is necessary to get back to where you want to be.  A recent reply to my previous blog post described medications as "tools" rather than a panacea and I think that until perfectly safe and effective medications are invented that is true.  Healy's point is that the advertising notion of "Take an antidepressant and get better" is false.  Psychiatrists are trained to help you navigate the complicated process of recovery from depression and side effects and the potential for side effects is generally the most complicated aspect.

George Dawson, MD, DFAPA


Additional Clinical Note 1:  Another blogger sent me an e-mail earlier this week asking me to send a list of psychiatrists who I thought were competent to taper people off of SSRI/SNRI type antidepressants.  The intention of the e-mail was to have a ready list of people who could help people with that particular problem.  I think that all psychiatric residents should be taught about medication discontinuation effects and how to resolve them, but apparently that is not the experience of some people who end up taking these medications.  As an instructor in a psychopharmacology course, I can verify that the residents I taught were all aware of this problem and how to deal with it.  They also had the very good back up reference of the ASCP psychopharmacology course PowerPoints and lecture materials on this problem.  I realize that this blog is not widely read, but I would appreciate any posts from instructors or professors about the issue of side effect recognition and treatment in general and SSRI discontinuation symptoms in particular and the approach to teaching these topics in your program.  I would also appreciate hearing your thoughts on this problem about SSRI/SNRI discontinuation symptoms and the variable experiences of people trying to get the problem diagnosed and treated.

Additional Clinical Note 2:  The processes that I am describing in the above post take time.  In many cases the equivalent amount of time required to do psychotherapy and longer.  I do have people telling me that their physicians (all types) seem to be poised over a prescription pad.  They tell me nobody has ever informed them of the risks or potential side effects of a medication.  I don't think the problem has been investigated and it would be difficult to do.  The idea that "medication management" in psychiatry, internal medicine or any other field is a brief uncomplicated encounter that takes little thinking on the part of a physician is largely an invention of business interests seeking to reimburse physicians at the lowest possible rate.  If you are a consumer of medical services, consider my approach in the above post and ask yourself if you have had the discussions that I describe.