# Lesson 3: Pharmacology of Psychedelics: Receptors, Pharmacokinetics, Dose and Interactions

*Companion-podcast transcript • Sarah & Kiffer*  
*Office Hours episode to listen to after working through the lesson*

---

**Sarah:** Here is a timing problem. The free plasma half-life of psilocin, the molecule that does the work after a dose of psilocybin, is under two hours. In the Basel studies, the experiences lasted between five and six hours.

**Kiffer:** So by the time a person is halfway through the afternoon, most of the drug that started the experience has already been cleared, and they are still fully in it. Any model that reads the plasma concentration as a thermometer for the experience gets this badly wrong.

**Sarah:** Which raises the obvious question. If the concentration in the blood is falling away, what is the experience actually tracking?

**Kiffer:** Something that has been measured directly in people, and the measurement turns into arithmetic you can do yourself. That is where I would like to start today.

**Sarah:** Welcome back to Office Hours. I'm Sarah.

**Kiffer:** And I'm Kiffer. That timing problem sits at the centre of Lesson three, Pharmacology of Psychedelics: Receptors, Pharmacokinetics, Dose and Interactions, so let's work it through.

**Sarah:** Question one. Why does the experience outlast the drug?

**Kiffer:** Because of the shape of the occupancy curve. Madsen and colleagues scanned eight healthy volunteers with positron emission tomography after oral psilocybin and fitted a single-site binding model. Occupancy equals a fitted maximum of 76.6 percent, times the plasma psilocin concentration, divided by the sum of that concentration and 1.95 micrograms per litre. The 1.95 is the concentration that reaches half of the attainable maximum. Everything interesting follows from where the curve bends.

**Sarah:** Let's make everyone do one. Take a plasma psilocin concentration of 5 micrograms per litre. Using those two fitted values, what share of receptors is occupied, and is it above or below half of the maximum? Take a few seconds, and pause the audio if you'd like more time.

*(Pause)*

**Kiffer:** Put 5 into the model. 76.6 times 5 is 383. The denominator is 1.95 plus 5, which is 6.95. And 383 divided by 6.95 is about fifty-five percent, so already well above half of the fitted maximum of 76.6 percent.

**Sarah:** Let me do the peak now, because this is the number from a dose people actually read about in trial reports. Holze and colleagues measured a mean maximal psilocin concentration of 17 nanograms per millilitre after a 25 mg dose. So I convert to micrograms per litre by dividing by a thousand, which gives 0.017. Then 76.6 times 0.017 is about 1.3, divided by 1.967, and I get under one percent occupancy.

**Kiffer:** Hold that answer up beside the measurements before you go on. What did the scans show?

**Sarah:** Between forty-three and seventy-two percent. Under one percent is impossible. I converted when I should not have. A nanogram per millilitre and a microgram per litre are the same quantity, so 17 nanograms per millilitre is 17 micrograms per litre.

**Kiffer:** Good catch, and that unit pair catches people constantly, because the pharmacokinetic papers and the imaging papers report in different units for the same quantity. Try the calculation again with 17.

**Sarah:** 76.6 times 17 is about 1302, and the denominator is 18.95. That gives about sixty-nine percent. So going from 5 to 17, more than a threefold rise in concentration, moves occupancy from about fifty-five to about sixty-nine percent.

**Kiffer:** Here is the harder version, and it is worth doing with a pen. Turn the model around. What concentration would you need for seventy percent occupancy, and what concentration for seventy-five percent? Take a few seconds.

*(Pause)*

**Sarah:** Rearranged, the concentration equals the occupancy you want times 1.95, divided by 76.6 minus the occupancy you want. For seventy percent, that is 136.5 divided by 6.6, which is about 21 micrograms per litre. For seventy-five percent, it is 146.25 divided by 1.6, which is about 91.

**Kiffer:** So five more percentage points of occupancy costs more than four times the concentration. That is the practical meaning of saturation at the receptor, and it is worth keeping separate from the subjective dose-response, which was still climbing across the three doses Holze and colleagues compared.

**Sarah:** There are two traps in using that model as a calculator, though, and I want to name both. The first is that 91 micrograms per litre is roughly five times the highest concentration Madsen and colleagues actually measured, so that answer sits well outside the data the curve was fitted to. The second is that the same study fitted a separate curve for subjective intensity, and its half-maximal concentration is 4.5 micrograms per litre, more than twice the value for occupancy.

**Kiffer:** Taking the second of those, the two curves are not interchangeable. At about 91 micrograms per litre, occupancy has essentially stopped moving while predicted intensity is still climbing towards its own fitted ceiling of 10.8, which sits just above the top of the zero to ten rating scale. Someone who read only the occupancy curve would predict a plateau in experience that the intensity curve does not show.

**Sarah:** And the fit itself was weaker for intensity, wasn't it.

**Kiffer:** Considerably. Intensity tracked occupancy more closely than it tracked the plasma concentration, and the whole model was fitted in eight people with wide variation between individuals in the concentration reached from a given dose. The radioligand is itself an agonist, so the occupancy estimate is tied to the method. My view is that the curve is a good teaching tool for the shape of saturation and a poor predictor of what one person will experience.

**Sarah:** Question two. What does the time course actually decide?

**Kiffer:** Staffing, mostly, and that is less mundane than it sounds. The Basel group measured onset, peak and offset the same way across agents, so the numbers can be compared. Mean offsets were 5.3 hours for 20 mg of psilocybin, 8.6 hours for one hundred micrograms of LSD, and 12.0 hours for 500 mg of mescaline. MDMA at one hundred milligrams came in at 4.9 hours. A ketamine infusion over forty minutes is a different world, with dissociation scores back to an absent level within an hour of the end.

**Sarah:** Let's turn that into something a service would have to budget for. Take a hypothetical outpatient service with two staff present for the whole session and two hours of preparation and discharge around it. Over forty completed sessions in a year, how many more staff hours does an LSD session cost than a psilocybin session? Take a few seconds.

*(Pause)*

**Sarah:** Psilocybin is 5.3 hours plus two, which is 7.3, times two staff, so 14.6 staff hours. LSD is 8.6 plus two, which is 10.6, times two, so 21.2. The difference is 6.6 staff hours per session, and over forty sessions that is 264 staff hours, roughly seven weeks of one person's full-time hours.

**Kiffer:** And mescaline breaks the model altogether. Twelve hours of acute effects plus the same two hours is a fourteen-hour day, which runs well past an ordinary eight-hour shift for a two-person team. The reported offset range for the higher mescaline dose ran from 7.9 to 22 hours, so a service planning from the mean is planning for a person who may not show up.

**Sarah:** That range is the part I would put in front of a program director. The mean is a scheduling number, and the tail is a staffing risk, because the person whose effects run to twenty-two hours is the one who decides whether a service needs overnight cover. It also changes the question a team asks about an agent. The useful question is which agent fits the hours a service can actually staff with two trained people present, and that pushes towards the shorter-acting agents for reasons that have nothing to do with how well any of them works.

**Kiffer:** Agreed, and it is worth saying what sets these differences. Mescaline outlasts LSD even though their elimination half-lives are close, because it reaches its maximal concentration later and holds a broader peak. The clinic day is set by the time to peak as much as by clearance.

**Sarah:** Question three. A news article says that psychedelics are non-addictive and therefore safe. What is wrong with that sentence?

**Kiffer:** Two separate things, joined by a word that does no work. The dependence evidence is reasonably consistent. Heal and colleagues, reviewing the abuse potential literature in 2026, report that self-administration experiments with classic psychedelics have not shown efficacy, and they conclude that most, and possibly all, conventional serotonin 2A agonist psychedelics will not serve as positive reinforcers in that model.

**Sarah:** They also give two explanations for the negative findings, which is the part people skip.

**Kiffer:** They do. Animals may not experience the sensory and perceptual changes as rewarding, or rapid tolerance may extinguish responding before a stable pattern can form. Those explanations matter, because a negative result in a model constrains a claim less firmly than a positive one. The model detects reinforcement reliably for stimulants and opioids, which is what makes a negative finding informative at all.

**Sarah:** And MDMA sits in a different place.

**Kiffer:** It does. MDMA maintains self-administration in rhesus monkeys, which puts it with drugs that function as reinforcers. Grouping it with the classic psychedelics under one heading about addiction potential hides a real pharmacological difference.

**Sarah:** There is one piece of the tolerance evidence I think gets underused clinically, which is cross-tolerance. Isbell and colleagues showed in 1961 that tolerance built to one of these drugs carries over to the other, in both directions.

**Kiffer:** That finding did two jobs. It was behavioural evidence for a shared mechanism, made before anyone had identified the receptor, and it agrees with the blockade studies. It also gives a clinician something concrete. When a person describes a recent experience that seemed unexpectedly weak, repeated use in the preceding days is one ordinary explanation, alongside the dose, the preparation and the setting.

**Sarah:** Then there is the second half of the sentence, the word therefore.

**Kiffer:** Which is where I would push hardest. A statement about reinforcement carries no information about cardiovascular risk, psychiatric risk, or the hazards of an unsupervised setting. Those are the risks that screening is built to manage, and Lesson six takes them up. I think the honest summary is that these drugs rarely produce compulsive use and that this tells you very little about whether any particular person should take one.

**Sarah:** Before question four, a short content note, because this part describes deaths reported after a drug interaction and after ibogaine use. If you would like support while listening, you can call or text nine eight eight, the Suicide Crisis Helpline, at any time, and the Hope for Wellness Help Line for Indigenous people is available at one eight five five, two four two, three three one zero.

**Kiffer:** And one boundary, since we are about to talk about combinations. Everything here is about clinical screening and about reading the research literature. Psilocybin, psilocin, LSD, mescaline and DMT sit in Schedule III of the Controlled Drugs and Substances Act, and MDMA and ketamine in Schedule I, so any session in Canada happens through an authorized route, which Lesson eleven covers.

**Sarah:** Question four. What is a medication review actually looking for?

**Kiffer:** Three mechanisms, and sorting a medication into the right one is most of the work. Does the other drug change how much agent reaches the circulation, usually through a liver enzyme. Does it occupy the same receptor or transporter. Or does it push the same physiological system in the same direction, whether that is serotonin activity, heart rhythm or blood pressure.

**Sarah:** Let's work a hypothetical referral that is not in the reading. Invented person, invented list. Someone is being considered for MDMA-assisted therapy through an authorized route, and they take duloxetine for depression and tramadol as needed for back pain. Which of those two would most change the team's thinking, and by which mechanism? Take a few seconds.

*(Pause)*

**Sarah:** Duloxetine is the interesting one, because it does two things at once. It blocks the serotonin and noradrenaline transporters, and MDMA works by entering the terminal through those transporters and reversing them, so the response may be blunted. Hysek and colleagues showed exactly that: duloxetine pretreatment prevented the rise in well-being and emotional excitation. It is also named among the antidepressants that inhibit the enzyme CYP2D6, which would push exposure up.

**Kiffer:** Two mechanisms pointing opposite ways in one medication, which is why a screening note cannot be written as a single direction of risk. On which one would most change the decision, I would give that to the tramadol.

**Sarah:** Because it is a weak serotonin reuptake inhibitor.

**Kiffer:** Yes, and because of what it would be sitting alongside. MDMA releases serotonin, the duloxetine is already holding the reuptake transporters, and the tramadol adds a third serotonergic drug to the same day. Foong and colleagues, writing for Canadian family physicians, identify a monoamine oxidase inhibitor combined with a serotonin reuptake inhibitor as the most concerning route to serotonin toxicity, and they list tramadol, dextromethorphan and St John's wort among the drugs that raise serotonin. An as-needed painkiller and a winter cough preparation are the two items a history built around psychiatric medication will miss.

**Sarah:** And the serious end of this is documented.

**Kiffer:** It is. Vuori and colleagues described four deaths after MDMA was taken together with moclobemide, a reversible monoamine oxidase inhibitor, and concluded that serotonin toxicity from the interaction was the probable cause in each case. None of the four had been prescribed the moclobemide, and the authors called for warnings about attempts to potentiate the effects of unregulated drugs.

**Sarah:** The third mechanism, additive physiology, is easiest to see with ibogaine, which is covered in Lesson ten for opioid use disorder.

**Kiffer:** It is the clearest case, because the physiology and the exposure problem meet in one drug. Koenig and colleagues showed, in cells engineered to carry the channel, that concentrations in the range reached during treatment reduce currents through the potassium channel that governs the heart's electrical recovery, which is a mechanism for a dangerous arrhythmia. Their later review sets out the case reports of long QT intervals and sudden cardiac death, and asks clinicians to clarify electrolytes, existing cardiovascular disease, interactions and genetic variation in metabolism before ibogaine is given.

**Sarah:** And methadone is the medication they single out.

**Kiffer:** They do, and it is a hard combination to screen around, because methadone prolongs the QT interval, has a half-life of 15 to 55 hours, and inhibits the same enzyme that determines how much ibogaine exposure a person gets. Alper and colleagues reviewed the fatalities outside West Central Africa that were temporally related to ibogaine between 1990 and 2008. Nineteen people died, and in 12 of the 14 cases with adequate postmortem data, advanced pre-existing conditions, mainly cardiovascular, or other substances explained or contributed to the death.

**Sarah:** Which reads as a screening failure as much as a drug effect.

**Kiffer:** Partly, and I would hold that reading carefully. The review found no characteristic syndrome of neurotoxicity, and in most of the cases with postmortem data an advanced heart condition, or another substance taken around the treatment, explained or contributed to the death. The authors also set out several pathways by which ibogaine itself may cause or contribute to a death, so this is information about screening and about the drug. What a case series like this cannot give is a risk under medical supervision with cardiac monitoring, and Lesson ten takes up the supervised cohorts.

**Sarah:** The lithium finding is the one I find hardest to place, because the design is weak and the signal is alarming.

**Kiffer:** It is the best example in this area of evidence that is weak in form and still decision-relevant. Nayak and colleagues analysed reports posted on three websites. Of sixty-two describing a classic psychedelic taken with lithium, forty-seven percent involved seizures and thirty-nine percent involved medical attention. Of thirty-four lamotrigine reports, none described a seizure.

**Sarah:** With no denominator, no verification of what was actually taken, and a strong pull towards posting dramatic outcomes.

**Kiffer:** All true, and two of the authors are employed by the organization that runs one of the sources, which the paper discloses. What survives those criticisms is the internal comparison. Both sets of reports come from the same websites and the same posting behaviour, so reporting bias has to explain why seizures appear in one group and not the other. My position is that this is enough to shape a screening conversation and nowhere near enough to put a rate on.

**Sarah:** Okay. Now the part where we disagree, which is what to do about antidepressants before a psilocybin session.

**Kiffer:** My default would be to continue the antidepressant unless the prescriber has a reason to change it. Becker and colleagues ran a double-blind crossover study in which escitalopram pretreatment left the positive mood effects of psilocybin intact and significantly reduced bad drug effects, anxiety and adverse cardiovascular effects, with psilocin kinetics unchanged. Goodwin and colleagues then gave a single 25 mg dose to nineteen people with treatment-resistant depression who stayed on their reuptake inhibitor, and the mean depression score fell by 14.9 points at three weeks.

**Sarah:** The Goodwin study was open-label with no comparison group, though, so it shows feasibility and cannot measure effect.

**Kiffer:** That is fair, and I would not lean on it alone. The controlled study is the stronger piece, and it points the same way.

**Sarah:** My worry comes from the survey. Barbut Siva and colleagues found that people who were not taking a reuptake inhibitor reported mystical experiences stronger by 18.2 percent, challenging experiences stronger by about fifty-one percent and emotional breakthrough stronger by about thirty-two percent. If the strength of the acute experience is associated with outcome, and Lesson five examines that evidence, then a blunted experience may be a weaker treatment.

**Kiffer:** A survey of people dosing outside a laboratory cannot separate the medication from the people taking it, though. Those two groups differ in illness severity, in history and in setting.

**Sarah:** Agreed, and the controlled study measured mood and adverse effects over one session in healthy volunteers, so it does not settle the therapeutic question either. What it establishes is that the positive acute effects survive escitalopram, which is a narrower claim than the one a team needs when it is deciding how to treat a person with depression. I am less comfortable than you are treating continuation as the default, mostly because the outcome that would settle this, symptom change at three or six months under each approach, has not been compared in a controlled trial.

**Kiffer:** Where I think we land together is on the process. The decision belongs to the prescriber who manages the medication, with the person and the team in the conversation, and stopping an antidepressant carries its own risk of symptoms returning during the weeks of preparation.

**Sarah:** And whichever way it goes, write it down properly. A monitoring plan during any taper, an agreed point at which the taper would be abandoned, and a plain account for the person of what is known and what is not.

**Kiffer:** Let's pull it together with three things to take away.

**Sarah:** First, saturation explains the timing. Occupancy rises steeply at low concentrations and flattens early, so an experience continues while the concentration falls, and five more points of occupancy can cost four times the concentration.

**Kiffer:** Second, a measured offset is a service design number. Two staff, two hours around the session and forty sessions a year separated psilocybin from LSD by 264 staff hours, and the reported range around a mean matters more than the mean.

**Sarah:** And third, name the mechanism and the strength of the evidence together. A controlled crossover study, an open-label series, a survey and a set of self-selected online reports support different sentences, and using one of them well means saying which it is.

**Kiffer:** For practice, rework the occupancy inversion from this episode. Find the concentration needed for sixty percent occupancy, then for seventy-two percent, and say in one sentence what the two answers show about dose.

**Sarah:** Next time, it's Lesson four, How Psychedelics Change the Brain: Networks, Plasticity and Learning, where we move downstream from the receptor to networks, plasticity and what animal findings can support.

**Kiffer:** Take care, everyone.

**Sarah:** See you in Lesson four.
