Who Should Never Take Ibogaine

Every clinic has a list. The lists are all the same list, and knowing where it came from tells you which parts to take seriously.

One document, copied forward for thirty years

The chain is documented and each link can be checked.

1993. A government drug institute drafted a protocol for a rising-dose tolerance study of ibogaine. It never produced a published dose-finding result.

2003. A treatment manual reproduced that protocol’s exclusion list almost verbatim, and included the original as an appendix. It is candid about the inheritance, and about the objection: one contributing author is quoted saying that by excluding patients who are depressed or bipolar you exclude a sizeable portion of the addict population.

2016. A provider association restructured the list into absolute exclusions and risk factors, and added several items. These are the guidelines most clinics cite.

2025. Two registered trials reproduced that 2016 absolute list word for word, in the same order, including its distinctive phrasing about bipolar disorder for which the patient has been hospitalised or medicated.

So an unrefereed document, which disclaims being a standard in its own text, has propagated into registered research protocols. How to evaluate a clinic covers what happened to the organisation that published it.

The exclusions with real evidence behind them

Treat these as absolute.

An existing prolonged QT interval. In a Dutch hospital study, maximum QTc prolongation averaged 95 milliseconds and half the patients exceeded 500 milliseconds. Starting from a prolonged interval means starting closer to the edge. The thresholds used are 450 milliseconds for men and 470 for women.

Electrolyte disturbance. In every ibogaine fatality where potassium was measured it was low, in one case as low as 2 millimoles per litre, and magnesium was low in half. This is the cheapest real risk reduction available, because it is correctable in an afternoon.

Structural heart disease. In the fatality series, advanced pre-existing conditions, mainly cardiovascular, explained or contributed to death in 12 of the 14 cases with adequate post-mortem data. Note the denominator: twelve of fourteen, not twelve of nineteen. Five cases had inadequate data.

Drugs that prolong the QT interval or inhibit CYP2D6. Both mechanisms are established in humans. Ibogaine drug interactions sets out which and why.

Liver impairment, though for a reason nobody wrote down at the time. See below.

The exclusions that are convention

These may still be prudent. They are not evidence-based, and it is worth knowing which is which.

Kidney disease. We searched the literature for ibogaine with renal injury and found nothing. There is no published case of ibogaine-associated kidney injury. The exclusion is a direct copy of a 1993 clause covering any disease of the gastrointestinal system, liver or kidneys that could alter metabolism or excretion, which is generic phase-1 boilerplate. The specific thresholds added later cite no source.

Liver disease. Also boilerplate, and also with no published hepatotoxicity case. But this one turns out to be well founded for the opposite reason from the one it was written for: ibogaine is cleared almost entirely by one liver enzyme, so the liver does not suffer from ibogaine, it controls how much of it reaches you. The right exclusion, for the wrong reason, by accident.

Pregnancy and breastfeeding. There is no evidence of any kind. No human data, no animal reproductive toxicity study. The exclusion appears first in 2003, is absolute from 2016, and cites nothing. Breastfeeding is not mentioned in either document and appears in only two registrations. Excluding pregnancy is a sensible default in the absence of data, and we say so rather than dressing it as a finding.

Age limits. Four registrations set four different upper caps for the same drug: 55, 60, 65 and 70. The provider guidelines set no cap at all, only a suggestion of a stress ECG above 60. Meanwhile every death in the fatality series was aged between 24 and 54, which is inside every one of those windows. The caps are trial-design convention, not a risk boundary.

One place the guidelines contradict the literature

The 2016 guidelines state that ibogaine itself has not been known to induce seizures, and frame the concern as belonging to alcohol and benzodiazepine withdrawal.

Three published sources record convulsions after ibogaine. A woman who took two grams had four to five seizure-like episodes with an apnoeic period, then QT prolongation and several episodes of torsades. A case report describes generalised tonic-clonic seizures after a very large dose. And a United Kingdom poisons-service series lists convulsions among the most frequent features across its seven patients.

The fatality series adds a screening domain the cardiac literature omits entirely: seizures associated with withdrawal from alcohol and benzodiazepines are named as an apparent risk factor.

The part where the field does not follow its own rules

The two most-cited outcome studies applied almost none of this.

A series of 191 patients states its entire exclusion list in one sentence: histories of stroke, epilepsy and psychotic disorders, cardiovascular and liver pathology, and HIV. A study of 30 patients at two Mexican clinics contains the words bipolar, psychosis, schizophrenia and contraindication zero times, and nine of its thirty patients had a previous psychiatric hospitalisation. A New Zealand study, in the one country where ibogaine can lawfully be prescribed, records that providers determined medical eligibility themselves. One of its fourteen patients died during treatment.

So the field states an exhaustive exclusion list and its own published cohorts did not apply it.

What this means for you

If any of the well-evidenced exclusions applies to you, the question is settled and no provider should proceed. If one of the conventional ones applies, you are being excluded by a document from 1993, which may still be the right call.

And if none applies, that is not a clearance either. Eight published cases involved people with no pre-existing cardiac condition at all, and a normal screening result misses a great deal.