The pharmacokinetics and pharmacodynamics of ibogaine in opioid use disorder patients

Ibogaine is an alkaloid from the rootbark of Tabernanthe iboga historically used in unregulated settings to treat addiction, including opioid use disorder (OUD). Earlier, non-controlled studies suggested ibogaine can mitigate opioid withdrawal and reduce craving and relapse, but dosing regimens have varied widely and no formal dose–effect studies exist. Major safety concerns remain, in particular ibogaine- and noribogaine-associated prolongation of the corrected QT interval (QTc), which raises the risk of torsades de pointes, and cerebellar ataxia observed in animal and human reports. Metabolism of ibogaine to noribogaine is largely mediated by cytochrome P450 2D6 (CYP2D6), with subsequent glucuronidation to noribogaine glucuronide (NIG); interindividual genetic variation in CYP2D6 activity could therefore drive exposure differences and safety risk, but the impact of genotype on PK and exposure–response relationships had not been fully characterised. Knuijver and colleagues designed a pharmacokinetic–pharmacodynamic (PKPD) study to address these gaps. Their primary aims were to quantify how CYP2D6 genetic variation affects ibogaine, noribogaine and NIG pharmacokinetics after a single oral dose, and to explore exposure–response relationships between plasma levels and three pharmacodynamic outcomes: QTc prolongation, cerebellar ataxia (assessed by SARA), and opioid withdrawal severity (OOWS and SOWS). The study was conducted in patients with OUD undergoing detoxification and monitored intensively for 24 hours after dosing.

This was an open-label PKPD study in 14 patients with opioid use disorder who were on agonist therapy (methadone or buprenorphine) and undergoing inpatient detoxification. Key exclusion criteria included clinically significant cardiac disease (history of ventricular fibrillation, long QT syndrome, syncope, or baseline QTc > 450 ms for men and > 470 ms for women), abnormal serum potassium (>5.0 or <3.5 mmol/l), severe liver or renal dysfunction, pregnancy, current use of QT-prolonging or CYP2D6-affecting medications (with the exception of methadone prior to inclusion) and a history of major depressive or psychotic symptoms. Participants received a single oral dose of ibogaine hydrochloride at 10 mg/kg administered in a yoghurt mixture at 08:30 after conversion to oral morphine and cessation of other opioid agonists to standardise withdrawal onset. Twenty milligrams of metoclopramide was given pre-dose to reduce nausea and secure ingestion. Subjects were admitted for eight days; the last morphine dose was given at 04:00 on day 9 and ibogaine was given approximately 4 hours later. Blood sampling included a pre-dose sample at 30 minutes before administration and further samples up to 24 hours post-dose; the extracted text does not clearly report all intermediate sampling times. ECGs were recorded every 30 minutes for the first 12 hours and then less frequently per protocol, with hourly measurements if QTc exceeded defined thresholds. If QTc exceeded 500 ms within the first 24 hours, magnesium infusion was administered for myocardial stabilisation. SARA (Scale for the Assessment and Rating of Ataxia), OOWS (Objective Opioid Withdrawal Scale) and SOWS (Subjective Opioid Withdrawal Scale) were assessed at 2, 6, 10 and 24 hours post-dose, except when subjects returned to opioid substitution therapy. Plasma concentrations of ibogaine, noribogaine and NIG were measured using a validated ultra-performance liquid chromatography–tandem mass spectrometry method (linear ranges reported for each analyte). CYP2D6 genotyping was performed by TaqMan analysis and genotypes were converted to an activity score (AS) for covariate testing. QTc was calculated using Fridericia's correction and QT intervals were averaged from leads V5 and II with verification by an independent cardiologist. Population PK analysis employed nonlinear mixed effects modelling (NONMEM 7.4) with a sequential approach: an integrated PK model for parent and metabolites was developed, CYP2D6 AS was tested as a covariate on formation/clearance parameters, and empirical Bayes estimates (EBEs) of individual concentration–time metrics (Cmax, Tmax, AUC) were derived. Exposure–response exploration used individual hysteresis plots and Spearman rank correlations; where associations were significant, nonlinear mixed effects PKPD modelling was applied (details reported in supplemental material).

Fourteen participants were included. Observed pharmacokinetics of ibogaine and its metabolites were highly variable across individuals; detailed Cmax, Tmax and AUC estimates are reported in the supplemental tables referenced in the text. Genotyping results per individual were also presented in the supplement. Population PK modelling identified a strong and statistically significant relationship between CYP2D6 activity score and the clearance of ibogaine to noribogaine (p < 0.0001). The model estimated a basic clearance of 0.82 L/h at an AS of 0, with an increase of 30.7 L/h for each additional point in AS. Sex was not identified as a significant covariate for the PKs of ibogaine or its metabolites. A total of 386 QTc measurements were available for analysis. Spearman correlation analysis found a modest but statistically significant association between ibogaine plasma concentrations and QTc (reported as r = 0.109, p < 0.05). Visual inspection of hysteresis plots showed no hysteresis for ibogaine in the PKPD section, while noribogaine and NIG exhibited clockwise hysteresis; the Discussion later describes an anticlockwise hysteresis for QTc plotted against ibogaine and clockwise hysteresis for noribogaine (the extracted text therefore contains an inconsistency on this point). Nonlinear PKPD modelling indicated that a sigmoid Emax relationship best described ibogaine concentration versus QTc prolongation, implying a plateau in effect. The model estimated a maximum QTc prolongation (Emax) of 67.9 ms (relative standard error 10.9%) and an EC50 of 0.195 µM (RSE 64.1%). The majority of observed ibogaine concentrations exceeded that EC50, consistent with rapid attainment of the plateau. Cerebellar ataxia assessed by SARA showed a strong Spearman correlation with ibogaine plasma concentrations (r = 0.67, p < 0.01), while correlations with noribogaine and NIG were not significant. For opioid withdrawal measures, 54 combined OOWS and SOWS assessments were performed and no significant correlations were observed between plasma levels (parent or metabolites) and withdrawal severity. The authors note that limited frequency and number of PD measurements might reduce sensitivity for detecting hysteresis or delayed effects.

Knuijver and colleagues interpret their findings as demonstrating that CYP2D6 genotype is a major determinant of ibogaine pharmacokinetics: reduced CYP2D6 activity substantially decreases clearance and therefore increases systemic exposure. Because ibogaine plasma concentrations correlated with QTc duration and the PKPD model predicts a concentration–dependent QTc effect that plateaus, the investigators conclude that higher exposure due to lower CYP2D6 activity is likely to increase both the magnitude and duration of QTc prolongation. They therefore suggest that CYP2D6-guided or otherwise individualised dosing should be considered to achieve safer and more consistent exposures. In terms of which compound drives observed adverse effects, the authors argue that cardiac (QTc) and cerebellar (ataxia) effects were most likely driven by parent ibogaine rather than noribogaine or NIG. Evidence supporting this interpretation includes the correlation of QTc and SARA with ibogaine concentrations but not with metabolites, the hysteresis patterns reported for metabolites, and prior animal data showing noribogaine produces less ataxia. Nonetheless, the authors acknowledge that because parent and metabolites coexist after dosing they cannot fully disentangle separate causal contributions in this dataset and that noribogaine at higher concentrations reported elsewhere may still impact QTc. No relationship was found between plasma levels of ibogaine or metabolites and objective or subjective withdrawal scores in these measurements. The investigators suggest this may reflect that the administered dose produced systemic exposures on the upper plateau of any concentration–effect curve for mitigation of withdrawal, so lower exposures might still be effective while reducing cardiac risk; however, the minimum effective exposure for withdrawal relief remains unknown. The authors acknowledge several limitations that affect interpretation and generalisability. The sample size was small (14 participants), although similar to other clinical ibogaine studies, and this limits precision and the ability to detect some PKPD relationships. Reliable QTc measurement is challenging and the study did not include a full day of QTc recording pre-dose to establish individual baseline circadian variation. The number and timing of SARA, OOWS and SOWS assessments were limited, which may have reduced sensitivity to detect delayed or transient effects. Potential pharmacokinetic interactions were noted: metoclopramide was administered pre-dose and in vitro data suggest it can be a reversible CYP2D6 inhibitor at high concentrations, but the presence of a clear CYP2D6 AS–PK relationship argues against complete inhibition in this study. Methadone and metoclopramide are listed as QT-prolonging agents and could have contributed to QTc duration, although baseline QTc values were within normal range and the temporal QTc prolongation followed ibogaine administration. Tobacco smoking, common in the population studied and permitted during the trial, may slightly influence QTc but the investigators did not observe changes in cigarette consumption or elevated baseline QTc that would suggest a major confounding effect. Overall, the authors recommend exploration of substantially lower doses and/or genotype-guided dosing strategies in future studies to improve cardiac safety while investigating whether reduced exposures retain clinical efficacy for withdrawal, craving or relapse prevention. They also note that noribogaine itself may warrant further clinical development given a potentially lower cardiotoxic profile, but additional data are required to separate parent and metabolite effects.