Ibogaine Success Rate: Why There Isn't One

Somebody quoted you a number, and it probably came with a confident tone.

This page is about where such numbers come from, which is a more useful thing to know than the numbers themselves.

What a success rate would require

Three things, none of which exists for ibogaine.

A denominator. Everyone who was treated, not everyone who answered afterwards. The people who died, who became too unwell to reply, or who never wanted to hear from the clinic again all belong in the count.

A definition fixed in advance. Abstinence from what, measured how, for how long, starting when.

A follow-up period that does not move. Reporting whoever happens to be reachable at whatever moment produces a figure that shifts with the effort put into finding people.

Every published ibogaine outcome figure is missing at least one of these. Most are missing all three.

The number everyone quotes

The 70 per cent figure comes from a retrospective survey of 88 people treated at one clinic in Mexico between 2012 and 2015.

Now the denominator above the denominator. The clinic had treated 336 people in that window. 285 could be contacted, 134 started the survey, and 88 remained after exclusions.

That is 26 per cent of the people treated. The authors say plainly that those who could not be contacted or declined may have had different outcomes, which may have inflated their estimates. Three of the five authors worked for the clinic, its owner among them, and the clinic funded the ethics application.

The same people, two headlines

A second paper analysed 73 of those same 88 respondents and reported that 59 of them, 81 per cent, were treatment responders.

So one dataset yields “70 per cent relapsed” and “81 per cent responded”. Both are accurate. A responder was anyone abstinent or using less than before; a relapse was any return to use at any point.

A clinic quoting the second figure and a critic quoting the first are describing the same eighty-eight people. Whenever you see an ibogaine success figure, the definition is doing more work than the number.

The best prospective evidence

Surveys ask people to remember. A prospective study follows them forward, which is harder and much more informative.

Thirty people treated in Baja California were followed for a year. The same cohort carries the withdrawal result, which is a stronger finding than this one and belongs beside it.

Improvement was greatest at one month and, in the authors’ words, never again reached that level. Twelve of the thirty met their definition of a favourable outcome at nine or twelve months.

Look at the middle column. The count of people actually reached moves up and down, because which people could be found changed each round.

The authors ran the comparison both ways, and it matters which one is quoted. Against where people started, the improvement stayed significant throughout. Against the first month’s own standard, their non-inferiority test returned no significant result at any later point once missing data were accounted for. Their own reading is that a subset responded and kept responding, rather than that the group as a whole did.

A New Zealand cohort of fifteen found roughly half using opioids again at three, six and twelve months. Its three dropouts had all relapsed. One participant died during treatment.

Why the biggest series cannot help

The largest medically monitored series, 191 patients in St Kitts, is often cited as evidence of effectiveness.

It reports no drug use outcome at all. What it measured was mood, craving and physician-rated withdrawal, and at one month those measures existed for roughly a third of the participants. Its author declares that she founded and holds shares in a company developing ibogaine and holds patents on its active metabolite.

It is a genuine safety observation. It is not a success rate and was never designed to be one.

What the whole literature amounts to

A systematic review counted 24 studies covering 705 people: seventeen open-label studies or case series, three case reports, one retrospective survey, and three controlled studies, none of which tested ibogaine for efficacy in dependence. Two deaths occurred within those 24 studies.

That is the entire human evidence base from which success rates are quoted.

The comparison that actually decides something

People weighing this are usually not choosing between ibogaine and nothing. They are choosing between ibogaine and staying on an opioid agonist, and that comparison has been measured, on the other side.

Across 749,634 people, all-cause mortality during agonist treatment was less than half what it was out of treatment. In the four weeks after stopping, all-cause mortality ran about six times higher than during treatment, and stayed around double for as long as people remained off it.

The comparison has never been run directly. Iboga compared sets out what is known about each alternative separately.

If someone quotes you a figure

Four questions dismantle almost any of them.

1. Out of how many treated? Not how many answered.
2. Success defined as what? Abstinence, or reduced use, or getting through withdrawal.
3. Measured when, and by whom? Self-report to the clinic that treated you is not a neutral measurement.
4. Who is missing? In every study here, the people not reached are the ones whose outcomes would move the number most.

The honest answer to “does it work” is that detoxification appears to work in the short term, that a minority of people in small studies were doing better a year later, that nobody has counted the rest, and that two randomised trials finished in 2024 without reporting.

When it does not work covers what the follow-up record says about the people for whom it did not.