This study designed an infusion protocol for DMT using pharmacokinetic/pharmacodynamic modelling. Comparing a continuous variable model to two bounded integer models, optimal doses for desired psychedelic intensity were identified. However, achieving consistent target intensity was challenging across models, indicating individual dose adjustments might be needed. Differences between models were especially notable at the scale's boundaries.
N,N-dimethyltryptamine (DMT) is a psychedelic compound that is being studied as a therapeutic option in various psychiatric disorders. Due to its short half-life, continuous infusion of DMT has been proposed to extend the psychedelic experience and potential therapeutic effects. The primary aim of this work was to design an infusion protocol for DMT based on a desired level of psychedelic intensity using population pharmacokinetic/pharmacodynamic modeling. As a secondary aim, the impact of choosing a continuous variable or a bounded integer pharmacokinetic/pharmacodynamic model to inform such an infusion protocol was investigated. A previously published continuous variable model and two newly developed bounded integer models were used to assess optimal doses for achieving a target response. Simulations were performed to identify an optimal combination of a bolus dose and an infusion rate. Based on the simulations, optimal doses to achieve intensity ratings between 7 and 9 (possible range = 0-10) were a bolus dose of 16 mg DMT fumarate followed by an infusion rate of 1.4 mg/min based on the continuous variable model and 14 mg with 1.2 mg/min for the two bounded integer models. However, the proportion within target was low (<53%) for all models, indicating that individual dose adjustments would be necessary. Furthermore, some differences between the models were observed. The bounded integer models generally predicted lower proportions within a target of 7-9 with higher proportions exceeding target compared with the continuous variable model. However, results varied depending on target response with the major differences observed at the boundaries of the scale.
Papers cited by this study that are also in Blossom
Sanches, R. F., Osório, F. L., Dos Santos, R. G. et al. · Journal of Clinical Psychopharmacology (2016)
D’Souza, D. C., Syed, S. A., Safi-Aghdam, H. et al. · Neuropsychopharmacology (2022)
Yaden, D. B., Griffiths, R. R. · ACS Pharmacology and Translational Science (2020)
Timmermann, C., Roseman, L., Schartner, M. et al. · Scientific Reports (2019)
Interest in new treatments for psychiatric disorders has renewed investigation of psychedelic compounds, including N,N-dimethyltryptamine (DMT). Previous population pharmacokinetic/pharmacodynamic work linked plasma DMT concentrations to a minute-by-minute subjective intensity rating on a discrete 0–10 scale and estimated an EC50 of about 92 nM. Most clinical studies to date have used intravenous bolus dosing, but because DMT has a very short half-life, continuous infusion has been proposed to extend the psychedelic experience and thus could be clinically useful. A methodological issue is that the commonly used continuous-variable pharmacodynamic models treat the 0–10 intensity score as effectively continuous and unbounded, whereas the data are discrete and bounded; bounded integer models have been proposed to respect those properties and may fit such data better. Eckernäs and colleagues set out to design an infusion protocol for DMT that targets a specified psychedelic intensity, and to examine whether choosing a continuous-variable versus a bounded-integer pharmacodynamic model materially affects dose selection. The work uses a previously published continuous-variable model together with two newly developed bounded integer variants (one including a Markov element for serial correlation) to simulate bolus-plus-infusion regimens and identify combinations that keep simulated subjects within predefined target intensity ranges over an extended period.
The analyses used data from a previously published, placebo-controlled clinical study at Imperial College London in which 13 healthy subjects (placebo on first visit, DMT on second) received intravenous bolus DMT fumarate at four dose levels. Plasma samples were taken up to 60 minutes post-dose and subjective intensity ratings (0–10) were recorded every minute for the first 20 minutes. The investigators analysed the data using nonlinear mixed-effects modelling in NONMEM v7.4.3 with Pirana and Perl-speaks-NONMEM for automation and R for diagnostics and plots. Individual pharmacokinetic parameters were estimated simultaneously with pharmacodynamic parameters, and effect-compartment models were used to describe the slight delay between plasma and effect-site concentrations. Model comparison used changes in objective function value (ΔOFV) with a −3.84 threshold for p = 0.05, and parameter precision was assessed by sampling importance resampling (5000/1000). The continuous-variable model was the previously published effect-compartment Emax (sigmoid Emax) model with E0 fixed to 0.001 and a logit transformation applied to keep predictions within the 0–10 boundaries on the transformed scale. For the bounded integer approach, the 11-category scale was modelled by dividing the standard normal distribution into 11 equal-area intervals via probit cutoffs; a normal mean and variance were modelled as functions of time, covariates and random effects, and the probability of each integer score was calculated from the cumulative normal distribution. Between-subject variability (BSV) was included on slope (drug effect) and on the variance function; a Markov element for serial correlation was also evaluated and implemented in one bounded-integer variant. Simulations incorporated BSV and residual variability to evaluate optimal bolus-plus-infusion combinations that would maintain steady ratings over an extended period (target considered over ≈60 minutes). Two targets were evaluated: a higher target of ratings 7–9 (considered desirable here) and a lower target of 4–6 for model comparison. Simulations used 1000 virtual subjects; peak effect-compartment concentrations were expected near 2 minutes after bolus so that timepoint was used for early assessment, and steady-state was defined as a time when more than five terminal plasma half-lives had elapsed in all simulated individuals. For mapping continuous-model outputs to integer targets, the investigators treated continuous predictions ≥6.5 and <9.5 as equivalent to 7–9, and ≥3.5 and <6.5 as 4–6. Once optimal doses were proposed, typical-subject time-course simulations (no variability) illustrated expected ratings over 60 minutes.
Model development and diagnostics indicated that a linear concentration–effect function best described the bounded integer data, with BSV on slope and on the variance term; including BSV in the variance improved fit. Adding a Markov element produced a substantial objective function improvement (ΔOFV = −83), but the authors also present a bounded-integer model without the Markov term to allow comparison with the continuous model that lacks a serial-correlation element. Visual predictive checks (VPCs) and parameter precision suggested broadly similar fits across the continuous-variable and bounded-integer models; individual pharmacokinetic parameter estimates were judged similar across models. Simulations identified different optimal bolus-plus-infusion combinations depending on the model and the target. For the target intensity 7–9, the continuous-variable model predicted the highest proportion within target with a 16 mg DMT fumarate bolus followed by an infusion of 1.4 mg/min. The bounded-integer models identified 14 mg bolus plus 1.2 mg/min infusion as the corresponding optimal combination. At those optimal combinations the proportion of the simulated population within the 7–9 target was low overall: about 45% for the continuous-variable model and 24%–26% for the two bounded-integer models, indicating substantial residual variability and the need for individual dose adjustments. When considering median responses across dose ranges, the continuous model predicted median 7–9 ratings for bolus doses 12–26 mg and infusion rates 1.2–2.6 mg/min, whereas the bounded-integer models produced narrower intervals (bolus 12–16 mg, or 14–18 mg with the Markov term; infusion 1.0–1.4 mg/min). At the high end of dosing the models diverged markedly in predicted extreme scores. The bounded-integer models tended to predict a much larger proportion of subjects attaining the maximum score of 10 at high doses (up to 94% reporting 10 at the highest simulated dose) compared with the continuous model (about 52% at the highest dose). For the lower target 4–6, all models converged on an optimal bolus of 10 mg with an infusion of 0.8 mg/min and similarly low proportions within target (<40%), again indicating that individually tailored dosing would be required. Typical-subject simulations (no variability) showed that the proposed bolus-plus-infusion combinations could maintain steady ratings over 60 minutes in a hypothetical typical individual.
The investigators used both a previously published continuous-variable pharmacodynamic model and two newly developed bounded-integer models to propose infusion protocols intended to achieve target subjective intensity ratings for DMT over an extended period. Eckernäs and colleagues report that, although the optimal bolus-plus-infusion combinations were similar between modelling approaches (16 mg + 1.4 mg/min for the continuous model versus 14 mg + 1.2 mg/min for the bounded-integer models to target 7–9), the predicted distribution of responses differed importantly. The bounded-integer models forecasted lower proportions within the 7–9 target at the proposed optimal doses and a higher frequency of maximum (10) ratings at high doses, whereas the continuous model predicted a higher proportion within target and fewer subjects exceeding the target range. The authors attribute these discrepancies principally to the bounded integer model’s respect for the discrete scale and boundaries, which changes behaviour near the scale endpoints. The study team notes that the Markov element improved fit to the observed data but had only minor influence on the dose-selection simulations. They emphasise that inter-individual variability is substantial: simulated proportions within target were generally low (<53% for the higher target and <40% for the lower target), implying that a single fixed regimen is unlikely to place a large fraction of subjects in the desired intensity window and that individualised titration will be needed. Potential covariates such as body weight, metabolic enzyme polymorphisms, or baseline neuropsychological factors could explain some variability, but the small sample size (13 subjects) precluded meaningful covariate analysis. Limitations acknowledged by the authors include reliance on a single small dataset drawn from subjects with prior psychedelic experience, which may underestimate intensity in the general population, and the fact that the recommended doses have not been clinically tested. They also point out modelling-structure differences that complicate a direct comparison (for example, BSV placed in the bounded-integer variance function versus in residual error for the continuous model), though sensitivity simulations without the variance BSV did not materially change the dosing conclusions. Visual predictive checks indicated that all models fitted the observed data comparably well, but the authors underline that model choice matters most near the boundaries of the rating scale. In conclusion, the paper presents dose recommendations intended as starting points for infusion studies, emphasises the need for individual dose adjustment, and highlights that bounded-integer versus continuous-variable model choice can materially influence predicted proportions above or at the maximum rating scale.
In recent years, the need for new treatments in psychiatric disorders, such as depression and anxiety, has led to an increase in research with different psychedelic compounds.The effects of these compounds naturally include a subjective component related to the psychedelic experience and it has been hypothesized that this component is closely related to therapeutic effects.Consequently, accurate characterization of the relationship between drug exposure and the subjective psychedelic experience may prove important to guide dose decisions in future clinical studies. The relationship between exposure to the psychedelic compound N,N-dimethyltryptamine (DMT) and its psychedelic effects has previously been investigated using population pharmacokinetic/pharmacodynamic modeling.The psychedelic effects were assessed through a subjective intensity score where individuals were asked to rate the intensity of the experience on a numbered scale from 0 to 10. Higher concentrations of DMT were associated with a higher rating. The concentration needed to achieve a rating of 5, corresponding to 50% of the maximum effect (EC 50 ), was estimated at 92 nM. The model was used to simulate the expected ratings associated with different intravenous bolus doses, demonstrating how this model can be used to guide dose decisions. However, whereas most published studies with DMT so far have used intravenous bolus doses to investigate its effects,a continuous infusion of DMT rather than bolus doses has been proposed as a suitable way forward to extend the psychedelic experience.Recently, infusions of DMT over varying lengths of time have been evaluated in the clinic.The subjective intensity score used in the aforementioned studyis a discrete outcome variable that can only assume integer values within certain boundaries. However, the published model treats the intensity ratings as being continuous which means that predictions of non-integer values can also occur. Another approach to handling this type of data, that was recently suggested by Wellhagen and colleagues,is the bounded integer model. This model respects both the discrete nature of the data and the boundaries of the scale. Previous work has shown that the bounded integer model is able to provide a better fit to the data than corresponding continuous variable models.However, limited information is available on how the two approaches compare in terms of applicability, for example, in a dosefinding context. The primary aim of this study was to design an infusion protocol for DMT based on a desired level of psychedelic intensity using the previously published continuous variable model as well as a newly developed bounded integer model. As a secondary aim, the impact of choosing one of the two different approaches on dose selection was investigated.
The data set used in this work was obtained from a previously published, placebo controlled clinical study performed at the Imperial College Clinical Research Facility, Imperial College London.DMT fumarate was administered as an intravenous bolus dose at four different dose levels to 13 healthy subjects. Each subject received placebo on their first visit and DMT on their second visit. Blood samples for quantification of DMT in plasma were collected up to 60 min after administration. The intensity of the subjective effects was assessed by asking subjects to rate the intensity of the experience on a scale from 0
This study aimed to design a dosing protocol for DMT based on a target psychedelic intensity. In addition, the impact of choosing a continuous variable or a bounded integer model to inform such a protocol was investigated.
Dose recommendations for DMT based on a target response is presented for the first time. It is also shown that no single dose is likely to lead to a large proportion of the population within target. Further, differences between the continuous variable model and the bounded integer model that may impact dose selection are demonstrated.
The dose recommendations presented here may impact the clinical development of DMT by providing a starting point for designing infusion studies with DMT.
OPTIMIZED INFUSION RATES FOR DMT to 10, where 0 is no effect and 10 is the most intense experience imaginable, every minute during the first 20 min after administration. The study was conducted according to the revised Declaration of Helsinki (2000), the International Committee on Harmonization Good Clinical Practices guidelines, and the UK National Health Service Research Governance Framework, and was approved by the National Research Ethics Committee London -Brent and the Health Research Authority. All subjects provided written informed consent to participate in the study.
A continuous variable model describing the relationship between DMT exposure and subjective psychedelic experience based on these data has already been published.In the present work, a bounded integer model was also developed in order to assess whether the choice of model might impact future dose decisions. The two models are described in more detail below. Data were analyzed using nonlinear mixed effects modeling in NONMEM version 7.4.3. (ICON Development Solutions).Pirana (version 3.0.0) and Perlspeaks-NONMEM (version 5.2.6)were used for model automation and diagnostics. R (version 4.1.1) was used for model diagnostics and visualization. Models were fitted using the first-order conditional estimation method with interaction or the Laplace estimation method for the continuous variable and bounded integer model, respectively. A population pharmacokinetic parameter and data approach was used, where the same population pharmacokinetic parameters were used as input in both models whereas individual pharmacokinetic parameters were estimated simultaneously with the pharmacodynamic parameters.Individual pharmacokinetic parameters were assessed to assure that there were no major differences between the models that might impact the interpretation of the results. Effect compartment models were used to describe the slight delay in response as compared to plasma concentrations.The change in concentration in the effect compartment is described according to: where k e0 is the effect compartment equilibrium rate constant, C p is the plasma concentration of DMT, and C e represents the theoretical concentration in the effect compartment. Model discrimination between nested models was based on objective function value (OFV) where a change in OFV of -3.84 was considered a significant model improvement at p = 0.05 under the assumption that ΔOFV is approximately χ 2 distributed. The fit of the models to the data was further assessed using visual predictive checks (VPCs). Sampling importance resampling (samples/resamples = 5000/1000) was performed to determine precision of the parameter estimates.
The continuous variable model was a previously published model. More details on model development and assessment can be found elsewhere.In brief, the relationship between DMT exposure and subjective intensity was described by an effect compartment model with a sigmoid maximum effect (E max ) response according to: where E 0 is the baseline response, E max is the maximum response, EC 50,e is the concentration of DMT at the effect site required to produce half of the maximum response, and the Hill coefficient γ describes the sigmoidicity of the relationship. For technical reasons, E 0 was fixed to 0.001, a more detailed discussion on this can be found in the original publication. However, for any practical purposes, 0.001 can be considered equal to 0 in this context. To keep predictions within the boundaries of the scale, a logit transformation was used for every observation j of each individual i as: where λ is the individual prediction, and ε ij is the residual error, additive on the logit scale, and following a normal distribution.
Given a scale with n categories, the area under the standard normal distribution is divided into n equally sized areas in the bounded integer model. This is done using the probit function to define n -1 cutoff values (Z 1/n -Z (n-1)/n ). Because the subjective intensity rating scale consists of 11 categories, the area under a standard normal distribution was divided into 11 equally sized areas. A function describing the mean of a normal distribution, using fixed effects (θ), random effects for an individual i (η i ), time, and covariates (X i ), f(θ,η i,f ,t,X i,f ), together with a function describing the variance of a normal distribution, using fixed effects (σ), random effects for an individual i (η i ), time, and covariates (X i ), g(σ,η i,g ,t,X i,g ), are used along with the cutoff values to estimate the probability of each score. The probability for the kth score is: where Φ if the cumulative distribution of the normal distribution function. For the first category (k = 1) this collapses into: and for the last category (k = n) into: representing the cumulative distribution within the intervals (-∞, Z 1/n ) and (Z (n-1)/n , ∞), respectively. Linear and power functions were evaluated to describe the relationship between DMT concentrations and the individual prediction of the mean of the normal distribution. Exemplified here by a linear relationship described as: where Base is the mean of the normal distribution before any drug is administrated, Slope is a constant describing the relationship between drug concentration and the individual prediction, and C e is the drug concentration in the effect compartment. Between subject variability (BSV) was assessed on drug effect and variance as exponential random effects following a log-normal distribution with mean zero and variance ω.In addition, a Markov element was implemented to evaluate any serial correlation in the data, as described by Wellhagen and colleagues.This was implemented as: where Y i,j-1 is the previous observation and P k,i,j is the probability of a score k for individual i at time j. If Y i,j and Y i,j-1 are different, the equation simplifies to: A positive value of the Markov parameter PM is associated with a higher probability of an observation having the same value as the previous observation. A more detailed description of the bounded integer model can be found elsewhere.
Simulations including both BSV and residual variability were performed to assess optimal dose levels for achieving different target intensity ratings. Doses were assessed as DMT fumarate doses throughout this work. The overall aim was to identify a combination of a bolus dose and an infusion rate that would keep subjects at steady ratings over a longer period of time (e.g., 60 min). The previously published continuous variable model and two newly developed bounded integer models were used to simulate subjective intensity ratings at 2 min after administration of a bolus dose as well as at steady-state in 1000 virtual subjects across different dose levels. A timepoint of 2 min after bolus administration was chosen, as it is the time where peak effect compartment concentrations would be expected based on the final models. Steady-state was defined as a timepoint where more than five terminal plasma drug half-lives had passed in all simulated subjects. For the purpose of this work, ratings between 7 and 9 were considered desirable. This was to ensure a strong psychedelic experience, as this has been hypothesized to correlate with therapeutic outcome,while also avoiding participants experiencing adverse psychological reactions (i.e., extreme anxiety). Consequently, ratings above 9 were considered undesirable in this context, whereas ratings below 7 were considered a subtherapeutic response. A lower target of ratings between 4 and 6 was also evaluated for the purpose of model comparison. For the continuous variable model, ratings greater than or equal to 6.5 and less than 9.5 would be considered equal to 7-9 and ratings greater than or equal to 3.5 and less than 6.5 would be considered equal to 4-6. For each evaluated dose level, the proportion of the population within, below, or exceeding target was simulated. Once optimal doses had been defined, simulations of ratings over time in a typical subject (i.e., without any variability), following administration of these doses were performed. For the bounded integer models, typical ratings were defined as the rating with the highest probability at each timepoint.
For the bounded integer model, a linear function best described the relationship between drug concentration and effect. BSV was incorporated on drug effect (slope) as well as the variance function (SD). Including variability in the variance function significantly improved the fit of the model to the data. Adding a Markov element provided a better fit to the data (ΔOFV = -83). However, because no element to account for serial correlation is present in the continuous variable model, two final bounded integer models are presented here, with and without a Markov element. Final parameter estimates and VPCs of all models, including the previously published continuous variable model, are presented in Tableand Figure, respectively. Based on the VPCs, the fit appears similar between models. The number of parameters, parameter precision, and the estimated size of the BSV is also similar between models. Individual pharmacokinetic parameter estimates from the different models were considered similar (data not shown). Residual scatterplots as well as distribution of individual pharmacodynamic parameter estimates are provided in Appendix S1. The model code is provided in Data S1. A simulated dataset is provided in Data S2.
The distributions of simulated subjective intensity ratings at 2 min after the administration of different bolus doses as well as at steady-state across different infusion rates with the different models are presented in Figure. For the bolus doses, simulated median ratings between 7 and 9 were achieved at dose levels ranging from 12 to 26 mg with the continuous variable model, whereas the corresponding dose ranges were 12-16 and 14-18 mg for the bounded integer model without and with a Markov element, respectively. At steady-state, a predicted median response between 7 and 9 was achieved at infusion rates of 1.2-2.6 mg/min with the continuous variable model and 1.0-1.4 mg/min for the two bounded integer models. Figuresandshow the percentage of individuals having an intensity rating within, above, or below target across different dose levels. Tables summarizing the predicted can be found in Appendix S1. The doses that are predicted to achieve the highest proportion of the population within target with the different models are summarized in Table. At a target of 7-9, the proportions within target at the optimal dose levels varied substantially between the different models whereas they are more similar at a target of 4-6. Figuredemonstrates the simulated typical ratings after administration of the optimal doses to achieve ratings between 7 and 9 presented in Table.
The primary aim of this work was to design a dosing protocol for achieving a target psychedelic intensity level over an extended period of time (e.g., 60 min) using a previously published continuous variable model as well as a newly developed bounded integer model describing the relationship between DMT plasma concentrations and ratings of the subjective intensity of the psychedelic experience. The data used in this work come from a previously published study where the psychedelic experience was assessed through a subjective rating score on a scale from 0 to 10.Two bounded integer models are presented here, with and without a Markov element accounting for serial correlation in the data. However, no clear impact of the Markov element on the simulations presented here was observed. Although the addition of a Markov element led to a slightly larger proportion below target at each bolus dose as compared to the bounded integer model without a Markov element, the influence on dose decisions has to be considered minor. Based on the results of the simulations, optimal dose levels to achieve target ratings between 7 and 9 or 4 and 6 were explored. Using the continuous variable model, we predict that the highest proportion within a target of 7-9 would be achieved with a bolus dose of 16 mg followed by an infusion of 1.4 mg/min. The corresponding combination for the bounded integer models was 14 mg combined with mg/min. These doses can be expected to provide ratings within target over an extended period of time in a typical subject. However, taking variability into account, the predicted proportions within target are generally low (<53%), indicating that individually adjusted doses are necessary to achieve a large proportion of the population within target. At a target response of 4-6, the highest proportion within target was achieved with a bolus dose of 10 mg combined with an infusion rate of 0.8 mg/min for all models. In addition, for this target, the proportions within target were low (<40%) and individually tailored doses are recommended. Nevertheless, whereas ratings above 9 were considered undesirable here, DMT is generally considered safeand there are no established adverse levels. Although ratings of 10 might imply intolerance, one cannot make that inference directly from these ratings alone. Further questions regarding, for example, anxiety would have to be asked to infer negative valence or intolerance with a rating of 10. Furthermore, bolus doses higher than the ones proposed here have been tested in the clinic without resulting in any safety concerns.Consequently, we believe that the proposed doses could serve as a good starting point. Potentially, one could then adjust the infusion rate gradually based on the reported ratings of each individual subject. There may also be underlying covariates, such as weight, polymorphisms in metabolizing enzymes, or baseline neuropsychological factors, driving the large variability observed here. Unfortunately, due to the low number of subjects in this study, no covariate analysis could be performed here. Future studies should Note: Doses are expressed as mg DMT fumarate. Abbreviation: DMT, N,N-dimethyltryptamine. T A B L E 2 Suggested optimal doses based on predicted proportions within the target response. Response is measured as a subjective intensity rating on a scale from 0 to 10.
Simulated typical rating over time after administration of a 16 mg DMT bolus dose followed by an infusion of 1.4 mg/min over 60 min for the continuous variable model and a 14 mg bolus dose followed by an infusion of 1.2 mg/min over 60 min for the bounded integer models. For the bounded integer models, the plotted typical rating corresponds to the rating with the highest simulated probability at each timepoint. DMT, N,N-dimethyltryptamine. focus on characterizing potential covariates to allow for individualized dosing based on such variables. A secondary aim of this work was to further investigate the impact of the model structure, that is, either a continuous variable model or a bounded integer model, on dose decision making. Interestingly, although the simulated optimal doses were similar between the different models, some key differences were identified. One major difference lies in the frequency of the predicted population falling within target as well as below or above target at the predicted optimal dose levels (Figuresand). At the predicted optimal infusion rates to achieve ratings between 7 and 9 of 1.4 and 1.2 mg/min for the continuous variable and the bounded integer models, respectively, the corresponding proportions within target are 45% for the continuous variable model and 24%-26% for the two bounded integer models. This difference is mainly due to a higher proportion of ratings exceeding the target of 9 based on the bounded integer models rather than any difference in the proportion of people at subtherapeutic levels. The bounded integer model thus indicates that a larger proportion of the population would need dose adjustments to reach therapeutic levels as compared to the continuous variable model. Furthermore, with a target level of ratings between 7 and 9, the continuous variable model predicts a larger proportion within target even at higher doses as compared to the bounded integer model. This can again mainly be attributed to a smaller proportion of the population above target with the continuous variable model. With the bounded integer model, we predict up to 94% of people having a rating of 10 at the highest simulated dose level as compared to 52% with the continuous variable model. Similarly, there is a difference in the predicted dose intervals leading to median ratings between 7 and 9. For the continuous variable model bolus doses between 12 and 26 mg as well as infusion rates between 1.2 and 2.6 mg/min are predicted to lead to median ratings within target. The corresponding dose intervals for the bounded integer models are 12-16 mg (or 14-18 mg when a Markov element is included) and 1.0-1.4 mg/min. In other words, the bounded integer model predicts a higher risk of people reporting ratings of 10, which could potentially mean an increased risk for adverse psychological reactions, at dose levels where only a small proportion of the population would be expected to report ratings of 10 based on the continuous variable model. This difference could have a major impact on decision making and consequently study outcomes in a clinical development setting. It should be noted that it is clear from the results of the simulations that the impact of choosing different modeling approaches depends on the target response and that the major difference between the two models is the behavior at the boundaries of the scale. Because the continuous variable model does not truly respect the boundaries or the discrete nature of the data this naturally affects the predictions. However, the impact is most prominent at the higher dose levels where the predicted score gradually gets closer to 10 without being able to reach an actual value of 10. If on the other hand, one is aiming for a medium intensity rating of between 4 and 6, the results from the two different models are very similar. The predicted optimal doses are identical between the models. However, the dose intervals leading to a median response within target varies slightly at bolus doses of 8-10 mg for the continuous variable model and bounded integer model without a Markov element compared to 8-12 mg when a Markov element is added. For the two bounded integer models, only an infusion rate of 0.8 mg/min is predicted to lead to a median response within target, whereas the corresponding range for the continuous variable model is 0.8-1.0 mg/min. Additionally, Figuresanddemonstrate that the predicted proportions of the population falling within a target of 4-6 are similar between the models, further demonstrating that the main difference lies at the boundaries of the scale. Whereas we believe that the dose recommendations provided here may serve as a good starting point when designing an infusion study, some limitations should be highlighted. First, this work focuses on a single dataset, based on only 13 subjects, and the models and hence dose recommendations may change with more data available. It should also be pointed out that the data which the models are built on are derived from subjects with previous experience of psychedelics. Consequently, there could be an underestimation of the intensity ratings if applied to the general population. Second, because the recommended doses have not yet been tested in the clinic, no conclusions can be made at this point regarding their clinical suitability. However, although a recent study, using doses similar to what has been reported here, observe slightly lower mean intensity ratings than what would be predicted based on the models presented here, the large variability observed also in that study further demonstrates that individualized dosing will likely be necessary.Further, this work was not focused on making the models fully comparable as such and that may have impacted the results. For example, the included BSV term in the variance function of the bounded integer models allows the consistency in ratings to vary between individuals, something that could be considered equivalent to including BSV in the residual error for the continuous model.However, to make sure that this had no major impact on the comparison between the models, simulations were also performed using a bounded integer model without BSV in the function. These simulations showed that, whereas this BSV term improved the fit of the model to the data, it had no major impact on the behavior of the model in the context of designing a dose protocol based on the simulations presented here (data not shown). Moreover, based on the VPCs presented in Figure, all models seem to fit the data well. More importantly, the fit appears similar between them. Nevertheless, there are some expected differences in the behavior of the models that may impact predictions. For example, for the continuous model, the baseline intensity score was fixed at 0. Whereas with the bounded integer model, based on the estimated base parameter and variability, scores above 0 are sometimes predicted even before any drug has been administered. This can also be observed in the simulations where the 90% prediction interval includes ratings of 1 when the administered dose is zero (Figure). Furthermore, because individual pharmacokinetic parameters were estimated for each model, this could cause differences in simulated outcomes. However, the aim was not to perform an extensive investigation on the behavior or appropriateness of the different models but rather to give dose recommendations for DMT specifically as well as to provide an example of how the choice of model might impact dose selection when planning a clinical study. To conclude, this study presents, for the first time, dose recommendations for DMT based on a target response level. Overall, it appears that the choice of optimal dose levels based on the target intensity would be similar regardless of model choice. However, it is clear that individual dose adjustments will be needed and that no single dose will lead to a high proportion of the population within the target range. Furthermore, the bounded integer and the continuous variable models do behave differently in terms of describing the variability. Hence, there are larger differences at target response levels approaching the boundaries of the rating scale. Dose decisions based on a continuous variable model may lead to a higher risk of observing maximum ratings of 10, whereas predictions based on the bounded integer model favor a more conservative approach in this context.
Create a free account to open full-text PDFs.
Gallimore, A. R., Strassman, R. J. · Frontiers in Pharmacology (2016)
Luan, L. X., Eckernäs, E., Ashton, M. et al. · Journal of Psychopharmacology (2023)