Assessing an ibogaine provider from the outside looks like a matter of impressions: how the website reads, how the staff sound, what former clients say. There is a better method available, and it is underused. The published studies describe their screening criteria, exclusions and monitoring arrangements in specific detail, which makes those protocols a documented benchmark that any provider can be measured against.
What follows is drawn entirely from the methods sections of papers in the peer-reviewed literature, including two studies conducted at private clinics rather than universities. The resulting questions are concrete and have right answers. A provider operating carefully will recognise them and answer readily; one that cannot answer has disclosed something important by that fact alone.
Where the benchmark comes from
Four protocols are useful. A 191-person inpatient series in St Kitts [1]. A Dutch study of 14 patients in a university psychiatry department [2] with a companion pharmacokinetic analysis [5]. The clinic protocol used for 30 veterans assessed at Stanford [3]. And a survey of 88 people treated at a residential clinic in Mexico, which documents that clinic's screening [4].
Those studies differ in design quality and none of them establishes that ibogaine is effective. That is not what they are being used for here. They are being used as a record of what competent practitioners considered necessary before giving a large single dose of a drug with a documented cardiac effect, and on that narrow question they converge closely.
The reasoning behind the benchmark is endorsed in review. A 2026 review notes that fatalities are documented in medically unsupervised settings and suggests prescreening and cardiac monitoring may reduce the risk [8]. Screening and monitoring are the two levers, and they are both observable in advance.
Screening before dosing
Every protocol includes an electrocardiogram. The St Kitts series performed a physical examination, safety laboratory tests and a 12-lead electrocardiogram [1]. The Mexican clinic took a history and physical, a 12-lead electrocardiogram, a blood count and a metabolic panel [4]. The clinic treating the veteran cohort required bloodwork and an electrocardiogram [3].
A 12-lead electrocardiogram rather than a single-lead reading is the standard, because it is what allows the QT interval to be measured properly. The measurement has a purpose beyond record-keeping: it establishes the baseline against which any prolongation during treatment is judged, and it identifies people whose interval is already too long to proceed safely.
Blood tests matter for a specific reason that is easy to miss. Potassium and magnesium affect cardiac repolarisation independently of any drug, and low levels raise arrhythmia risk. The Dutch protocol required potassium between 3.5 and 5.0 millimoles per litre before proceeding [2]. Someone arriving after a period of poor intake, vomiting or heavy drinking may have abnormal electrolytes, which is correctable but only if measured.
What the studies excluded
The exclusion criteria are where the protocols are most explicit, and they are the clearest test of whether a provider has thresholds or merely preferences. The St Kitts series excluded cardiovascular and liver pathology, a resting heart rate below 50 beats per minute, and long QT syndrome [1].
The Dutch study excluded a QT interval above 450 milliseconds in men or 470 in women, potassium outside the normal range, severe liver or kidney dysfunction, and any QT-prolonging or CYP2D6-affecting medication [2]. The Mexican clinic excluded prolonged QT interval and heart disease [4]. The veteran cohort's clinic excluded cardiovascular, liver or kidney history, psychosis and neurological disorder [3].
The common thread is numbers rather than judgements. A provider who states a specific QT threshold above which they will not dose is operating a criterion. One who says each case is assessed individually, without naming a figure, has no threshold, and a criterion that bends under commercial pressure is not a criterion. The question worth asking is simply what measurement would cause them to decline, and roughly how often that happens.
That last part is informative. A provider who has never turned anyone away either sees an unusually healthy population or is not applying the criteria they describe.
If something here speaks to your situation, the next step is a five-minute provider match profile. We read it personally, take your case to vetted providers, and come back with the ones who say yes — or tell you honestly if ibogaine is not the answer.
Find a providerMedication review and washout
The most thorough documented washout comes from the clinic treating the veteran cohort, which stopped CYP2D6 inhibitors, serotonergic drugs, calcium channel blockers, beta blockers, benzodiazepines, stimulants, corticosteroids and all psychiatric medications before treatment [3]. That breadth reflects an assumption that anything acting on the heart or on liver metabolism is a potential contributor.
Two things make this more than a formality. Blocking CYP2D6 roughly doubled exposure to ibogaine plus noribogaine in healthy volunteers [9], so an unidentified inhibitor meaningfully raises the dose that is effectively delivered. And methadone itself inhibits CYP2D6 and is likely to impair ibogaine metabolism [6], which is why the published series substituted morphine before dosing [1] [2].
Washout duration is the part most often handled poorly. Some drugs, including several antidepressants and their active metabolites, persist for weeks after the last dose, so a few days is inadequate for them. A provider who specifies different washout periods for different drug classes understands the pharmacology. One who gives a single blanket figure for everything does not.
Withdrawing that many medicines carries its own risks, and a careful provider will say so. Stopping psychiatric medication can destabilise the condition being treated, and abrupt benzodiazepine cessation is itself dangerous. A washout should be supervised, which means someone medically qualified managing it rather than a written instruction sent before travel.
The genotype question
This is the question that most reliably separates thorough practice from adequate practice. CYP2D6 genotype drove a more than tenfold spread in ibogaine clearance among patients given the same dose per kilogram, and QT prolongation and ataxia tracked ibogaine concentrations rather than noribogaine concentrations [5]. Slow metabolisers are therefore exposed to more of the compound that carries the cardiac risk, for longer.
The authors of the inhibitor study recommended that it may be prudent to genotype and to at least halve the intended dose in poor metabolisers [9]. The St Kitts series genotyped CYP2D6 in a subset of participants [1], so the practice exists outside purely academic settings.
Genotyping takes days to weeks depending on the laboratory, so it has to be arranged before travel rather than on arrival. A provider who raises it unprompted is unusually careful. One who has never heard of it is unaware of the single largest identified source of variation in how their own treatment affects people.
Monitoring during and after
Duration is the critical variable, and it is longer than intuition suggests. The St Kitts series ran electrocardiogram telemetry from one hour before dosing until 24 hours after [1]. The Dutch study monitored for at least 24 hours [2]. The veteran cohort was monitored for 72 hours [3].
The reason to prefer the longer end is documented. QT prolongation outlasts the presence of ibogaine in the blood [6], and in the Dutch cohort an interval above 450 milliseconds persisted beyond 24 hours in 6 of 14 patients [2]. Monitoring that ends when the psychoactive effects subside stops while the cardiac abnormality is still present in a substantial minority of people.
Continuous telemetry is also different from intermittent checks. An arrhythmia developing over minutes is caught by a monitor with an alarm and missed by someone taking a reading every few hours. The question is whether monitoring is continuous, what it covers, and for how many hours after dosing it continues.
Staff and equipment
The Mexican clinic documented in the survey had live cardiac monitoring, intravenous fluids and electrolytes, and physicians present [4]. Those three elements together are what makes a detected arrhythmia a treatable event rather than a fatal one.
This is where the New Zealand case is instructive. Treatment there was delivered by legal providers, and one participant died during treatment. The coroner found no significant cardiac pathology and judged the death very likely related to ibogaine and most probably to a cardiac arrhythmia, and the provider was found to have breached its duty of care [7]. Legality and good practice are separate things, and the absence of detectable heart disease did not protect the person who died.
Practical questions follow: who medically qualified is on site and awake throughout, what resuscitation equipment is present, whether magnesium is available and under what circumstances it is given, as the Dutch protocol gave it when the interval passed 500 milliseconds [2], and how far the nearest hospital with cardiac capability is.
The material itself
One further question concerns what is being administered. The published studies used purified ibogaine hydrochloride at a measured dose, 8 to 12 milligrams per kilogram in St Kitts [1], 10 in the Dutch study [2], and around 15 at the Mexican clinic [4]. Total alkaloid extract contains other iboga alkaloids in uncontrolled proportions, so a stated weight of extract does not correspond to a known quantity of ibogaine.
A provider should be able to say which preparation they use and produce a certificate of analysis for the batch. Without that, the dose calculation rests on an assumption about content, which undermines the only link between a careful weight-based calculation and the exposure that results.
What the answers tell you
None of this makes treatment safe, and no combination of answers should be read that way. The cardiac effect is a property of the drug, and even under the careful screening of the Dutch study half the participants passed 500 milliseconds [2]. What these questions establish is whether a provider is working to the standard that published practitioners considered necessary, or to something less.
The answers also reveal how a provider responds to being asked. Specific figures, named thresholds, differentiated washout periods and a willingness to say what would cause them to decline all indicate a protocol. Reassurance without numbers, impatience at the questions, or claims that screening removes the risk all indicate its absence.
It is worth adding that a provider's own published standards, where they exist, can be checked against the protocols above rather than taken at face value. The documents in this literature are freely readable, and the specific figures in them, a QT threshold, a potassium range, a monitoring duration, a washout list, are the same figures a careful operation will be able to produce on request, in writing, without needing to be pressed for them.
Frequently asked questions
What is the single most important question?
How long monitoring continues after dosing. QT prolongation outlasts the drug [6], and an interval above 450 milliseconds persisted past 24 hours in 6 of 14 closely screened patients [2]. Published protocols ran from 24 to 72 hours [1] [3].
Is an electrocardiogram enough on its own?
No. Protocols pair it with blood tests, because potassium and magnesium affect arrhythmia risk independently and the Dutch study required potassium within a defined range before proceeding [2].
Does legal operation indicate quality?
Not by itself. In New Zealand, treatment by legal providers produced a death during a research study, and the provider was found to have breached its duty of care [7].
Should a provider ask about medications?
Yes, exhaustively. Blocking CYP2D6 roughly doubled active exposure in volunteers [9], and the most thorough documented protocol withdrew eight categories of medicine before dosing [3].
References
- Mash DC, Duque L, Page B, Allen-Ferdinand K. 2018, Front Pharmacol 9:529. frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2018.00529/full
- Knuijver T, Schellekens A, Belgers M, et al. 2022, Addiction 117(1). doi:10.1111/add.15448. PMC9292417. europepmc.org/article/PMC/PMC9292417
- Cherian KN, Keynan JN, Anker L, et al. 2024, Nat Med 30(2):373-381. doi:10.1038/s41591-023-02705-w. PMID 38182784. PMC10878970. europepmc.org/article/PMC/PMC10878970
- Davis AK, Barsuglia JP, Windham-Herman AM, Lynch M, Polanco M. 2017, J Psychedelic Stud 1(2):65-73. doi:10.1556/2054.01.2017.009. PMID 30272050. PMC6157925. europepmc.org/article/PMC/PMC6157925
- Knuijver T, ter Heine R, Schellekens AFA, et al. 2024, J Psychopharmacol 38(5):481-488. doi:10.1177/02698811241237873. PMID 38519421. PMC11102648. europepmc.org/article/PMC/PMC11102648
- Koenig X, Hilber K. 2015, Molecules 20(2):2208-2228. doi:10.3390/molecules20022208. PMC4382526. mdpi.com/1420-3049/20/2/2208
- Noller GE, Frampton CM, Yazar-Klosinski B. 2018, Am J Drug Alcohol Abuse 44(1):37-46. doi.org/10.1080/00952990.2017.1310218
- Shen MR, Owusu-Boaitey K, Murphy ZD, et al. 2026, Front Psychiatry 17:1909149. doi:10.3389/fpsyt.2026.1909149. PMID 42755825. PMC13581858. frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2026.1909149/full
- Glue P, Winter H, Garbe K, et al. 2015, J Clin Pharmacol 55(6):680-687. doi:10.1002/jcph.471. PMID 25651476. doi.org/10.1002/jcph.471