A Semiparametric Deconvolution Model to Establish In Vivo–In Vitro Correlation Applied to OROS Oxybutynin

Two-step in vitro-in vivo correlations: Deconvolution and convolution methods, which one gives the best predictability? Comparison with one-step approach

Finding predictive dissolution tests and valid IVIVCs are essential activities in generic industry, as they can be used as substitutes of human bioequivalence studies. IVIVCs can be developed by two different strategies: a one-step approach or a two-step approach. The objectives of this work were to compare different deconvolution and convolution methods used in the development of two-step level A IVIVCs and to study if the relationship between the in vitro dissolution rate and the in vivo dissolution rate should guide the decision between using a two-step approach or a one-step approach during the development of a new IVIVC. When the in vitro and the in vivo dissolution rates had a linear relationship, valid and biopredictive two-step IVIVCs were obtained, although there was not a combination of deconvolution and convolution methods that could be named as the best one, as long as all the prediction errors for any combination were within the limits. It was not possible to obtain a valid two-step IVIVC when the relationship between dissolution rates was non-linear, but the one-step approach was able to overcome this fact and it gave valid IVIVCs regardless of whether the relationship between dissolution rates was linear or non-linear.

Comparison of Alternative Population Modeling Approaches for Implementing a Level A IVIVC and for Assessing the Time-Scaling Factor Using Deconvolution and Convolution-Based Methods

Different approaches based on deconvolution and convolution analyses have been proposed to establish IVIVC. A new implementation of the convolution-based model was used to evaluate the time-scaled IVIVC using the convolution (method 1) and the deconvolution-based (method 2) approaches. With the deconvolution-based approach, time-scaling was detected and estimated using Levy's plots while with the convolution-based approach, time-scaling was directly determined by a time-scaling sub-model of the convolution integral model by nonlinear regression. The objectives of this study were (i) to show how time-scaled deconvolution and convolution-based approaches can be implemented using population modeling approach using standard nonlinear mixed-effect modeling software such as NONMEM and R, and (ii) to compare the performances of the two methods for assessing IVIVC using complex in vivo drug release process. The impact of different PK scenarios (linear and nonlinear PK disposition models, and increasing levels of inter-individual variability (IIV) on in vivo drug release process) was considered. The performances of the methods were assessed by computing the prediction error (%PE) on C_max, AUC, and partial AUC values. The mean %PE values estimated with the two methods were compliant with the IVIVC validation criteria. However, different from convolution-based, deconvolution-based approach showed that (i) the increase of IIV on in vivo drug release significantly affects the maximal %PE values of C_max leading to failure of IVIVC validation, and (ii) larger %PE values for C_max were associated to complex nonlinear PK disposition models. These results suggest that convolution-based approach could be considered at preferred approach for assessing time-scaled IVIVC.

Deconvolution Analysis by Non-linear Regression Using a Convolution-Based Model: Comparison of Nonparametric and Parametric Approaches

The convolution-based modeling approach has been shown to be flexible and easy to implement for performing a deconvolution analysis and for assessing in vitro/in vivo correlation using non-linear regression and a pre-specified model describing the in vivo drug absorption. A generalization of this method has been developed using a nonparametric description of the in vivo drug absorption process in replacement of a model-based definition. A comparison of the parametric and nonparametric deconvolution and convolution analyses was conducted on the pharmacokinetic (PK) data observed in four published studies after the administration of an extended-release formulation of methylphenidate at the dose of 18 mg. All the analyses were conducted using a conventional non-linear regression software (NONMEM). The results of the deconvolution analysis indicated that the parametric and nonparametric approaches performed similarly. The parametric approach described the input function using a double Weibull equation (6 parameters) while the nonparametric approach described the input function using a piecewise approximation (12-13 parameters). The validation of the results of the deconvolution analysis was conducted by comparing observed and predicted PK concentrations by the convolution analysis. The performance of the parametric and nonparametric approaches for assessing deconvolution was evaluated using the Akaike and the Bayesian information criteria. These criteria indicated that, despite the similar results obtained with the two approaches, the nonparametric approach provided better results. In conclusion, these results indicated that the nonparametric approach should be considered as the preferred approach for conducting a deconvolution analysis.

Development of a Novel Simplified PBPK Absorption Model to Explain the Higher Relative Bioavailability of the OROS® Formulation of Oxybutynin

A new minimal Segmented Transit and Absorption model (mSAT) model has been recently proposed and combined with intrinsic intestinal effective permeability (P _eff,int ) to predict the regional gastrointestinal (GI) absorption (f _abs ) of several drugs. Herein, this model was extended and applied for the prediction of oral bioavailability and pharmacokinetics of oxybutynin and its enantiomers to provide a mechanistic explanation of the higher relative bioavailability observed for oxybutynin's modified-release OROS® formulation compared to its immediate-release (IR) counterpart. The expansion of the model involved the incorporation of mechanistic equations for the prediction of release, transit, dissolution, permeation and first-pass metabolism. The predicted pharmacokinetics of oxybutynin enantiomers after oral administration for both the IR and OROS® formulations were in close agreement with the observed data. The predicted absolute bioavailability for the IR formulation was within 5% of the observed value, and the model adequately predicted the higher relative bioavailability observed for the OROS® formulation vs. the IR counterpart. From the model predictions, it can be noticed that the higher bioavailability observed for the OROS® formulation was mainly attributable to differences in the intestinal availability (F _G ) rather than due to a higher colonic f _abs , thus confirming previous hypotheses. The predicted f _abs was almost 70% lower for the OROS® formulation compared to the IR formulation, whereas the F _G was almost eightfold higher than in the IR formulation. These results provide further support to the hypothesis of an increased F _G as the main factor responsible for the higher bioavailability of oxybutynin's OROS® formulation vs. the IR.

Morphological changes in the salivary acini after in vivo cholinergic stimulation

AIMS

Overactive bladder syndrome treated by muscarinic receptor antagonists may be complicated by reduced salivation. Cholinergic agonists may reverse this effect. The aim of the present study was to determine the antagonizing effect of a cholinergic agonist (carbachol) on a muscarinic receptor antagonist (oxybutynin) in the submandibular acini in a rat model.

METHODS

Forty male Sprague Dawley rats were divided into three groups: Group I (control), Group II (vehicle), and Group III (treatment). Group III was subdivided so Group IIIa was treated with a muscarinic receptor antagonist (oxybutynin) for 1 week, Group IIIb was treated with oxybutynin for 3 weeks, and Group IIIc was treated with oxybutynin for 1 week and oxybutynin and a cholinergic agonist (carbachol) for 2 weeks. Histological and ultrastructural studies were performed on submandibular glands.

RESULTS

Group IIIa showed moderate atrophic changes in the serous acini and ducts. Group IIIb showed serous acini with distorted wall, widening of the inter-lobar space, and deposition of mononuclear cells in the connective tissue. Group IIIc had serous acini similar to Group I, with mildly dilated inter-lobar ducts, but some serous acini revealed double nuclei and the inter-lobar duct showed luminal vacuolations. Ultrastructural studies confirmed histological results.

CONCLUSIONS

Muscarinic receptor antagonist administration led to changes in the submandibular gland of rats, while concomitant administration of cholinergic agonists seemed to counteract these atrophic changes. Additional studies should assess carbachol as a cholinergic agonist in treating dry mouth in patients with overactive bladder syndrome who are taking the muscarinic receptor. Neurourol. Urodynam. 35:574-581, 2016. © 2015 Wiley Periodicals, Inc.

Oral osmotically driven systems: 30 years of development and clinical use

The number of marketed oral osmotically driven systems (OODS) has doubled in the last 10 years. The main clinical benefits of OODS are their ability to improve treatment tolerability and patient compliance. These advantages are mainly driven by the capacity to deliver drugs in a sustained manner, independent of the drug chemical properties, of the patient's physiological factors or concomitant food intake. However, access to these technologies has been restricted by the crowded patent landscape and manufacturing challenges. In this review article, we intend to give an overview of the OODS development in the last 30 years, detailing the technologies, specific products and their clinical use. General guidance on technology selection is described in light of the recent advances in the field. The clinical performance of these technologies is also discussed, with a focus on food effects and the in vivo-in vitro correlation. Special attention is paid to safety given the controversial case study of Osmosin. Overall, oral osmotically driven systems appear to be a promising technology for product life-cycle strategies.

Oral heroin in opioid-dependent patients: pharmacokinetic comparison of immediate and extended release tablets

In diacetylmorphine prescription programs for heavily dependent addicts, diacetylmorphine is usually administered intravenously, but this may not be possible due to venosclerosis or when heroin abuse had occurred via non-intravenous routes. Since up to 25% of patients administer diacetylmorphine orally, we characterised morphine absorption after single oral doses of immediate and extended release diacetylmorphine in 8 opioid addicts. Plasma concentrations were determined by liquid chromatography-mass spectrometry. Non-compartmental methods and deconvolution were applied for data analysis. Mean (+/-S.D.) immediate and extended release doses were 719+/-297 and 956+/-404 mg, with high absolute morphine bioavailabilities of 56-61%, respectively. Immediate release diacetylmorphine caused rapid morphine absorption, peaking at 10-15 min. Morphine absorption was considerably slower and more sustained for extended release diacetylmorphine, with only approximately 30% of maximal immediate release absorption being reached after 10 min and maintained for 3-4h, with no relevant food interaction. The relative extended to immediate release bioavailability was calculated to be 86% by non-compartmental analysis and 93% by deconvolution analysis. Thus, immediate and extended release diacetylmorphine produce the intended morphine exposures. Both are suitable for substitution treatments. Similar doses can be applied if used in combination or sequentially.

Estimation of the impact of noncompliance on pharmacokinetics: an analysis of the influence of dosing regimens

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

Patient compliance is better with formulations that require less frequent dosing than with formulations that require more frequent dosing. Intramuscular risperidone and long-acting oxybutynin are two examples of medicines reformulated for less frequent dosing. However, it is not clear whether better compliance with less frequent dosing regimens translates to improved therapeutic outcome.

WHAT THIS STUDY ADDS

At equivalent daily doses and typical patterns of compliance, fortnightly intramuscular depot administrations of risperidone provide better pharmacokinetic coverage than once-daily oral dosing. Once-daily dosing of oxybutynin is no better at maintaining pharmacokinetic exposure than twice-daily dosing at half strength. The use of simulated compliance data as input to pharmacokinetic models is useful to assess the impact of noncompliance on internal drug exposure.

AIMS

To determine whether, for oxybutynin and risperidone, drug exposure is better with less frequent dosing regimens than with regimens that require more frequent dosing.

METHODS

Pharmacokinetic models of oxybutynin (5 mg twice-daily and 10 mg once-daily) and risperidone (2 mg once-daily orally and 25 mg fortnightly intramuscular injection) were developed. Simulations of multiple doses were performed by use of stochastic models of dose-taking compliance and clinic visit attendance.

RESULTS

At therapeutic concentrations and with typical patterns of noncompliance, intramuscular injections of risperidone resulted in a 41% (SD 12%) greater pharmacokinetic coverage than the oral dose, 76% (SD 10%) vs. 35% (SD 7%). No discernable differences were evident between once- and twice-daily formulations of oxybutynin, 29.2% (SD 10%) vs. 29.0% (SD 13%).

CONCLUSIONS

For equivalent doses for each drug, the longer acting preparation of risperidone, but not oxybutynin, is pharmacokinetically more forgiving of noncompliance than the shorter acting counterparts. Further analysis is required to confirm whether these observations are valid clinically.

Level A in vitro-in vivo Correlation (IVIVC) Model with Bayesian Approach to Formulation Series

In vitro-in vivo correlation (IVIVC) models for formulation series are useful in drug development, but the current models are limited by their inability to include data variability in the predictions. Our goal was to develop a level A IVIVC model that provides predictions with probabilities. The Bayesian approach was used to describe uncertainty related to the model and the data. Three bioavailability studies of levosimendan were used to develop IVIVC model. Dissolution was tested at pH 5.8 with basket. The IVIVC model with Bayesian approach consisted of prior and observed data. All observed data were fitted to the one-compartment model together with prior data. Probability distributions of pharmacokinetic parameters and concentration time profiles were obtained. To test the external predictability of IVIVC model, only dissolution data of formulations E and F were used. The external predictability was good. The possibility to utilize all observed data when constructing IVIVC model, can be considered as a major strength of Bayesian approach. For levosimendan capsule data traditional IVIVC model was not predictable. The usefulness of IVIVC model with Bayesian approach was shown with our data, but the same approach can be used more widely for formulation optimization and for dissolution based biowaivers.

An automated process for building reliable and optimal in vitro/in vivo correlation models based on Monte Carlo simulations

Many mathematical models have been proposed for establishing an in vitro/in vivo correlation (IVIVC). The traditional IVIVC model building process consists of 5 steps: deconvolution, model fitting, convolution, prediction error evaluation, and cross-validation. This is a time-consuming process and typically a few models at most are tested for any given data set. The objectives of this work were to (1) propose a statistical tool to screen models for further development of an IVIVC, (2) evaluate the performance of each model under different circumstances, and (3) investigate the effectiveness of common statistical model selection criteria for choosing IVIVC models. A computer program was developed to explore which model(s) would be most likely to work well with a random variation from the original formulation. The process used Monte Carlo simulation techniques to build IVIVC models. Data-based model selection criteria (Akaike Information Criteria [AIC], R2) and the probability of passing the Food and Drug Administration "prediction error" requirement was calculated. To illustrate this approach, several real data sets representing a broad range of release profiles are used to illustrate the process and to demonstrate the advantages of this automated process over the traditional approach. The Hixson-Crowell and Weibull models were often preferred over the linear. When evaluating whether a Level A IVIVC model was possible, the model selection criteria AIC generally selected the best model. We believe that the approach we proposed may be a rapid tool to determine which IVIVC model (if any) is the most applicable.

A multi-mechanistic drug release approach in a bead dosage form and in vitro/in vivo correlations

An in vitro/in vivo relationship of a combined multi-mechanistic dosage form has now been established in the literature. In our previous study, we successfully prepared a combination of immediate release, enteric coated, and controlled-release (CR) beads and mathematically modeled in vitro and in vivo drug release characteristics of the combination based on the release profiles of individual beads. The objective of the present study is to develop in vitro/in vivo correlations (IVIVC) for three individual beads and the combination using theophylline as a model drug and the beagle dog as an animal model. In the study, an IVIVC correlation is estimated by two-stage procedures: deconvolution followed by comparison of the fraction of drug absorbed to the fraction of drug dissolved. The Wagner-Nelson mass balance method was used to deconvolute plasma drug concentration-time curves. In vitro, a two-stage medium (0.1 N HCl and pH 6.5 phosphate buffer) was used for the dissolution test; a 2h first stage (acidic) was selected based on the average gastric emptying time in a fasted dog. In vivo, t(lag) was used for the gastric emptying process for enteric coated beads and the combination, which contains enteric coated beads. A time-scaling technique was used to consider the rate difference between in vitro dissolution and in vivo absorption in the process of IVIVC. As shown in the results, a point-to-point correlation was established for each formulation. The linear regression analysis of the correlation was r2>0.99 for all three individual beads and 0.97 for the combined bead dosage form. The results suggest level A IVIVCs indicating an appropriateness for the in vitro and in vivo models used in this study.

Metabotropic glutamate receptor antagonist AIDA blocks induction of mossy fiber-CA3 LTP in vivo

Metabotropic glutamate receptors (mGluR) are implicated in long-term memory storage. mGluR-I and mGluR-II antagonists impede various forms of learning and long-term potentiation (LTP) in animals. Despite the evidence linking mGluR to learning mechanisms, their role in mossy fiber-CA3 long-term potentiation (LTP) is not yet clear. To explain the involvement of mGluR-I in memory mechanisms, we examined the function of the mGluR-I antagonist 1-aminoindan-1, 5-dicarboxylic acid (AIDA) on the induction of mossy fiber-CA3 LTP in vivo in male Sprague Dawley and Fischer 344 (F344) rats. Acute extracellular mossy fiber (MF) responses were evoked by stimulation of the MF bundle and recorded in the stratum lucidum of CA3. The excitatory postsynaptic potential (EPSP) magnitude was measured by using the initial slope of the field EPSP slope measured 2-3 ms after response onset. After collection of baseline MF-CA3 responses at 0.05 Hz, animals received either ((+/-))-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (N-methyl-D-aspartate-R antagonist, 10 mg/kg ip), naloxone (opioid-R antagonist, 10 mg/kg ip), or AIDA (mGluR antagonist, 1 mg/kg ip or 37.5 nmol ic). LTP was induced by two 100-Hz trains at the intensity sufficient to evoke 50% of the maximal response. Responses were collected for an additional 1 h. AIDA blocked induction of LTP in the mossy fiber pathway (P < 0.05) in both strains of rats after systemic and in Sprague Dawley rats after intrahippocampal injection.

Direct, differential-equation-based in-vitro-in-vivo correlation (IVIVC) method

A new, differential equation-based in-vitro-in-vivo correlation (IVIVC) method is proposed that directly relates the time-profiles of in-vitro dissolution rates and in-vivo plasma concentrations by using one- or multi-compartment pharmacokinetic models and a corresponding system of differential equations. The rate of in-vivo input is connected to the rate of in-vitro dissolution through a general functional dependency that allows for time scaling and time shifting. A multiplying factor that accounts for the variability of absorption conditions as the drug moves along is also incorporated. Two data sets incorporating slow-, medium-, and fast-release formulations were used to test the applicability of the method, and predictive powers were assessed with a leave-one-formulation-out approach. All fitted parameters had realistic values, and good or acceptable fits and predictions were obtained as measured by plasma concentration mean squared errors and percent AUC errors. Introduction of step-down functions that account for the transit of the dosage form past the intestinal sites of absorption proved useful. By avoiding the integral transforms used in the existing deconvolution- or convolution-based IVIVC models, the present method can provide increased transparency, improved performance, and greater modelling flexibility.

A computational model for particle size influence on drug absorption during controlled-release colonic delivery

The effect of particle size on the percent drug absorbed is computationally modeled for controlled-release dosage forms that deliver drug particles to the colon. The relative benefit of reducing particle size is mapped on a diagram of the drug's absorption rate constant (estimated from rat intestinal perfusion, CACO-2 or human intubation permeation rates) versus the drug's solubility. Some drugs fall into a limit of high percentage absorption even with large particles such that particle size reduction has little impact. Another group of drugs is solubility limited such that even with small particles, absorption is negligible. Between the two regions, only drugs with sufficiently high absorption rates are influenced by the drug dissolution rate and thereby the particle size. The size of this region is a function of dosing rate. Comparisons between calculated particle size effects on colon absorption as a function of colon volume suggest caution when using animal models to predict bioavailability from colonic drug delivery. This volume dependence also suggests that the particle size influence will vary as a function of the digestive cycle.

In vivo-in vitro correlation (IVIVC) modeling incorporating a convolution step

The purpose of in vivo-in vitro correlation (IVIVC) modeling is described. These models are usually fitted to deconvoluted data rather than the raw plasma drug concentration/time data. Such a two-stage analysis is undesirable because the deconvolution step is unstable and because the fitted model predicts the fraction of a dosage unit dissolved/absorbed in vivo which generally is not the primary focus of our attention. Interest usually centers on the plasma drug concentration or some function of it (e.g., AUC, Cmax). Incorporation of a convolution step into the model overcomes these difficulties. Odds, hazards, and reversed hazards models which include a convolution step are described. The identity model (which states that average in vivo and in vitro dissolution/time curves are coincident or directly superimposable) is a special case of these models. The odds model and the identity model were fitted to data sets for two different products using nonlinear mixed effects model fitting software. Results show that the odds model describes both data sets reasonably well and is a significantly better fit than the identity model in each case.