Is Iboga Safe? Ibogaine Risks and Screening

“Is iboga safe” is the most common way into this subject and the least useful question to ask, because it invites an answer of one word and the honest answer has a shape. That people have died after taking ibogaine, and that people have taken it in supervised settings and come through, are both in the literature, and a page that reports only one of them is selling something.

Iboga is the plant, ibogaine the alkaloid extracted from it, and almost everything measurable on this page was measured on the alkaloid. Root bark of unknown strength is a different exposure and a worse-documented one.

What the record actually contains

The most careful count comes from a systematic review of all available autopsy, toxicological and investigative reports, covering every known death outside West Central Africa temporally related to ibogaine between 1990 and 2008. It found nineteen individuals, fifteen men and four women aged 24 to 54, who died between 1.5 and 76 hours after taking ibogaine.

The exclusion is not incidental. Iboga root bark has been taken in Gabon and Cameroon for longer than anyone has counted, in quantities nobody has recorded, and no register exists that would let a death there enter a forensic series. The nineteen are deaths among people taking a purified or semi-purified alkaloid, mostly for addiction, mostly in Europe and North America. What the plant does in its own setting is not in this number and is not known.

That figure closed in 2008, and it is the one that circulates. A review chapter published in 2018 gave a detailed update of 33 deaths known to have occurred, five of them in the United Kingdom. The two counts use different geographies and different standards, so they cannot be subtracted from one another. A 2022 systematic review cross-checked the earlier series and found 38 deaths and 20 emergencies associated with ibogaine up to the end of 2020, again on its own basis. No count has been published since. Ibogaine deaths works through the whole sequence and explains why there is no death rate at all.

Two findings from the 2012 forensic review matter more than either number.

The first is what the postmortem evidence did not show. The authors looked for a characteristic syndrome of neurotoxicity, the sort of signature a drug leaves when it poisons the brain directly. They did not find one. Whatever is killing people is not a general toxic effect on nervous tissue.

The second is what the evidence did show. In 12 of the 14 cases where postmortem data were adequate, advanced pre-existing medical conditions, mainly cardiovascular, or one or more commonly abused substances, explained or contributed to the death. The review also identified seizures associated with withdrawal from alcohol and benzodiazepines, and the uninformed use of raw ethnopharmacological preparations, as further risk factors.

That is not a reassurance and it should not be read as one. It is a redirection. The question stops being “how dangerous is this drug in general” and becomes “what is already true of this person’s heart, and what else is in their body”. Those are answerable questions, which is the entire argument for screening.

Is ibogaine safe for the heart? Why cardiac risk is the whole argument

The mechanism is now reasonably well described. Ibogaine inhibits the hERG potassium channel at low micromolar concentrations, and hERG carries the current that repolarises the cardiac cell. Block it and repolarisation slows, the action potential lengthens, and the QT interval on the ECG stretches with it. A sufficiently long QT can degenerate into torsade de pointes, a ventricular arrhythmia that is sometimes self-limiting and sometimes fatal.

Published case reports bear this out at the bedside. QTc values above 600 milliseconds recur in the published reports of ibogaine-associated long QT, all of them survived cases; the fatalities have no recorded QTc. Against that, preclinical toxicology in dogs conducted on behalf of the US National Institute on Drug Abuse did not show QT prolongation, and no ECG abnormalities were reported under continuous monitoring in a series of 39 people treated for cocaine or heroin dependence, though the same review records significant bradycardia in six of those 39. Both are worth stating, because they are the kind of inconvenient result that tends to disappear from summaries in both directions.

One detail from the case reports deserves more attention than it usually gets. In the cases collected to 2015, low potassium was present in every fatality and arrhythmia case in which it was measured at all, and it was frequently not measured. Hypokalaemia independently lengthens the QT interval, and it is common in exactly the population that seeks ibogaine: people in withdrawal, vomiting, poorly nourished, sometimes drinking heavily. It is also trivially correctable with a blood test and repletion, though the published cases include potassium as low as 2 mM, half of them with low magnesium alongside it. A risk factor that is both prevalent and cheap to remove is the strongest available case for screening.

Ibogaine and the heart sets out the mechanism, why no QT threshold has been validated for this drug, and what monitoring is meant to catch.

Who should not take it at all

The exclusion lists are weaker than they look. Every list in circulation descends from a single drug-institute protocol drafted in 1993, copied into a 2003 manual and pasted into trial registrations since. QT prolongation, the mechanism behind most of the deaths, appears in none of the documents before 2016. For more than twenty years the thing that kills people was not an exclusion criterion.

Some exclusions are well founded and some are boilerplate nobody has tested. There is no published case of ibogaine causing kidney or liver injury, so those entries rest on convention. Four trial registrations set four different upper age limits for the same drug, and every death in the fatality series fell inside all of them.

What is well founded is cardiac history, and what is in the body alongside it. Drug interactions covers the drugs that inhibit the enzyme clearing ibogaine, which acts as a dose increase, and the drugs that lengthen the QT interval in their own right, which add to the mechanism directly. Coming off antidepressants covers the washout clinics require, which no published source has ever established a duration for and which is the least supervised part of the process. Bipolar disorder covers the psychiatric exclusion stated most often and examined least.

What a pre-treatment screen actually measures, and what it misses

Screening is the part a person can insist on, and it is worth knowing its limits before trusting it.

A baseline ECG with a calculated QTc, blood electrolytes with potassium and magnesium corrected before dosing, liver function, and a full list of every other drug the person takes: that is the floor. What screening should cover works through each.

The limits are real. In the study that defined congenital long QT, a resting ECG produced false negatives in forty per cent of male gene carriers. The American Heart Association states that monitoring the QT interval alone may be inadequate to predict the arrhythmia it is used to predict. In the best-monitored ibogaine study, two trained researchers measuring the same traces by hand differed by an average of thirty-four milliseconds, which is larger than most of the thresholds providers claim to work to. And which correction formula is used changes the answer, with Bazett over-reading at the heart rates typical of someone in withdrawal.

None of that is an argument against screening. It is an argument against treating a passed screen as a clearance, and it is why what should be in the room matters as much as what was measured beforehand.

The counter-evidence, and what it does and does not show

In 2024, a Stanford-led team, with co-authors from the treating clinic, published outcomes for 30 special operations veterans with traumatic brain injury who received ibogaine together with intravenous magnesium at a clinic in Mexico. Magnesium was given specifically to blunt QT prolongation. The study reported no bradycardia, no tachycardia, no haemodynamic instability and no QT prolongation detectable on monitoring, with QTc followed visually on continuous five-lead ECG rather than measured.

This result is real and it is regularly overstated. The authors say plainly that it was not a randomised controlled trial, and that participants had independently scheduled the treatment themselves, after grant approval from a veterans’ non-profit, before the researchers were involved. The authors describe the sample as highly homogeneous and mostly in above-average physical condition. That is not the population that appears in the fatality series.

The study is good evidence that this protocol produced no detectable rhythm disturbance and no haemodynamic instability in thirty people. It is not evidence that ibogaine is safe, and the paper does not claim it is.

Holding both of these at once is the honest position. The deaths are real, and so is their absence here. The two groups differ in more than one way, in baseline health, in monitoring, in the magnesium and in the dosing regime, and screening is the part a person can actually act on.

What happens outside a clinic

Everything above assumes a setting where something is measured. Much of what happens does not.

Iboga for sale online covers a supply chain with no cardiac screening at any point, and material whose alkaloid content nobody has established. The 2012 review named the uninformed use of ethnopharmacological forms as a risk factor in its own right, separate from the drug itself.

The sharpest case is a powder labelled Iboga du Gabon. A thirty-year-old woman died after taking it. Analysis found no ibogaine in it at all, and instead ajmaline, yohimbine and reserpine, which are alkaloids of Rauvolfia. Investigators had to compare the powder against an authentic iboga root sample to work out what it actually was.

One case does not establish how often that happens. It establishes something narrower and worse: a precaution about quantity protects nobody when the material is the wrong species.

What is still unknown about ibogaine safety

Several things that are routinely asserted are not established.

No QT threshold has been validated for this drug. Figures circulate, and screening protocols are built on them, but there is no QTc value above which ibogaine is agreed to be unsafe and below which it is agreed to be acceptable. The cut-offs providers use are house practice rather than validated limits, and the cardiac page sets out what that means for anyone being told they passed a screening. Nor has anyone separated the contributions of magnesium, electrolyte correction and monitoring, which are given together and have never been tested apart. It is reasonable to think screening helps. How much, and which part of it, is unknown.

There is no established safe dose. A Phase 1/2a study whose first stage was designed to determine the maximum tolerated dose, and which enrolled 116 people across its two stages and completed on 16 January 2024, has never posted results to the registry. Until it or something like it reports, no completed trial gives a basis on which a dose could be set, and the doses in the fatality series overlap the doses used in clinics. Anyone quoting a standard protocol is describing a house practice. Why there is no standard dose works through the four incompatible figures in circulation, and prints none of them as a recommendation.

The long-term picture is thin. A systematic review of the adverse-event literature published between 2015 and 2020 collects what has been published, but published series are dominated by short-term observation, and people who do badly months later are not systematically counted anywhere.

The denominator is unknown. Nineteen deaths between 1990 and 2008 is a rate only if you know how many people took ibogaine over that period, and nobody does. Most administration happens outside any registry. Fatality counts and per-treatment risk are different quantities and should not be presented as the same one.

The pages in this section

Every safety page below is written to be read on its own, and each carries a named reviewer and the date they read it.

Where a country stands, and what that permits, is a separate question answered on legal status by country.

We publish no supplier, no clinic recommendation and no affiliate link, here or anywhere on this site. Where a claim on these pages rests on a source we could not obtain, the page says so rather than rounding the gap into confidence.