Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Diabetes is one of the most common diseases in modern time. Its connection to overweight and obesity is well established, and diet and exercise are therefore important parameters in the treatment. A commonly used biomarker to diagnose and follow disease progression in diabetics is via measurements of fasting plasma glucose, FPG. In this study, the relationship between weight loss and FPG in overweight diabetics was studied. Competing hypothesis regarding the connection between weight loss and reduced FPG was investigated by using nonlinear mixed effects modeling based on data gathered from a meta-analysis by Anderson et al (1). The hypotheses suggested that either [1] weight effected FPG directly by an intermediate effector, or [2] both weight and FPG were affected by an unknown underlying mechanism. The intermediate effector was presumed to be insulin sensitivity and the underlying mechanism the blood concentration of free fatty acids.  The data was gathered from 8 different studies, all examining the results of very low energy diets (330-909 kcal/day) in overweight type 2 diabetics. Frequent measurements of weight and FPG were provided in each study with a range of 91-321 mg/dl for baseline FPG and 93-118 kg for baseline weight. The summarized studies consisted of 13 arms with 6-62 subjects in each arm. Both hypotheses were modeled by using NONMEM 7.2. A stepwise effect was used for both weight and FPG. For hypothesis [1], an inhibitory effect affected the weight input which then affected the output for insulin sensitivity by a relative change in weight or the input for the insulin sensitivity by an absolute weight change. For hypothesis [2] the same inhibitory effect affected weight input and the input for insulin sensitivity. For both models the FPG drop was then proportional to the increase in insulin sensitivity. Hypothesis [2] had a significantly lower objective function value (OFV) than hypothesis [1] and had also better results from goodness of fit plots and VPCs. It was therefore concluded that hypothesis [2] indicated the more accurate explanation of the connection between FPG and weight loss. Moreover, a strong correlation between the caloric content of the diet and the rate of weight change was seen as a result of stepwise covariate modeling. An impact from baseline BMI on rate of change for insulin sensitivity was also seen.

Modeling is a valuable tool in drug development, to support decision making, improving study design, and aid in regulatory approval and labeling. This thesis describes the development of pharmacometric models for drugs used in anesthesia and analgesia. Models describing the effects on anesthetic depth, measured by the bispectral index (BIS), for a commonly used anesthetic, propofol, and for a novel anesthetic, AZD3043, were developed. The propofol model consisted of two effect-site compartments, and could describe the effects of propofol when the rate of infusion is changed during treatment. AZD3043 had a high clearance and a low volume of distribution, leading to a short half-life. The distribution to the effect site was fast, and together with the short plasma half-life leading to a fast onset and offset of effects. It was also shown that BIS after AZD3043 treatment is related to the probability of unconsciousness similar to propofol. In analgesia studies dropout due to lack of efficacy is common. This dropout is not at random and needs to be taken into consideration in order to avoid bias. A model was developed describing the PK, pain intensity and dropout hazard for placebo, naproxen and a novel analgesic compound, naproxcinod, after removal of a wisdom tooth. The model provides an opportunity to describe the effects of other doses or formulations. Visual predictive checks created by simultaneous simulations of PI and dropout provided a good way of assessing the goodness of fit when there is informative dropout. The performance of non-linear mixed effects models in the presence of informative dropout, with and without including models that describe such informative dropout was investigated by simulations and re-estimations. When a dropout model was not included there was in general more bias. The bias increased with decreasing number of observations per subject, increasing placebo effect and increasing dropout rate. Bias was relatively unaffected by the number of subjects in the study. The bias had, in general, little effect on simulations of the underlying efficacy score, but a dropout model would still be needed in order to make realistic simulations.

Diabetes is a group of metabolic diseases characterized by hyperglycaemia resulting from defects in insulin secretion, insulin action, or both. Several models have been developed for describing the glucose-insulin system. Silber and Jauslin developed a semi-mechanistic integrated glucose insulin (IGI) model which simultaneously describe glucose and insulin profiles in either healthy subjects or type 2 diabetis mellitus (T2DM) patients. The model was developed for describing the basal system, i.e. when no drugs are present in the body. In this thesis the IGI model was extended to also include the effects of anti-diabetic drugs on glucose homeostasis. The model was extended to describe postprandial glucose and insulin excursions in T2DM patients treated with either biphasic insulin aspart or the GLP-1 receptor agonist liraglutide. These extensions make the model a useful tool in drug development as it can be used for elucidating the effects of new products as well as for clinical trial simulation. In this thesis several modelling tasks were also performed to get a more mechanistic description of the glucose-insulin system. A model was developed which describes the release of the incretin hormones glucosedependent insulinotropic polypeptide and glucagon-like peptide-1 following the ingestion of various glucose doses. The effects of these hormones on the beta cell function were incorporated in a model describing both the C-peptide and insulin concentrations in healthy subjects and T2DM patients during either an oral glucose tolerance test or an isoglycaemic intravenous glucose infusion. By including measurements of both C-peptide and insulin concentrations in the model it could also be used to characterize the hepatic extraction of insulin.

Artemisinin compounds are the most potent anti-malarial drugs available in the market. Today, malaria treatment is largely relies on the artemisinin-based combination therapies. Artesunate (AS) is the most widely used artemisinin derivative. In this thesis, we characterized the population pharmacokinetics of AS and its active metabolite dihydroartemisinin (DHA) following oral administration of AS in different populations. In Chapter II, we developed a population pharmacokinetic model of AS and DHA in healthy subjects. These subjects received either single- or multiple-dosing of oral AS, as a monotherapy regimen or in combination with pyronaridine, with or without food. In Chapter III, we developed a population pharmacokinetic model of AS and DHA in adult and pediatric patients with uncomplicated falciparum and vivax malaria who were administered oral pyronaridine/artesunate combination once daily for 3 days. We modeled the AS and DHA data simultaneously using a parent-metabolite model that assumed complete conversion of AS to DHA. Following oral administration, AS is rapidly absorbed with maximum concentrations reached at about 0.5 hours post-dose. AS is rapidly converted to DHA. DHA then undergoes rapid metabolism, with an elimination half-life of about 0.8 hours in malarial patients. Inter-individual variability for almost all pharmacokinetic parameters and residual variability for both compounds were estimated by the models. Substantial variability was seen in the pharmacokinetic parameters between the subjects. In healthy subjects, intake of food with the dose was found to delay the absorption of AS significantly, but not the extent of absorption. Weight was also included in this model as a determinant of DHA clearance. When modeling the data from patients, we included weight as part of the model a prioria priori using an established allometric function. No other covariates examined in the analysis were statistically significant. The performance of final models was evaluated using non-parametric bootstrap technique and visual predictive check. The models were found to adequately described the data at hand, and robust with sufficient predictive power. The results can be used as the base to develop a population pharmacokinetic-pharmacodynamic model and as prior information in guiding the selection of optimal sampling schedule for future pharmacokinetic studies of AS.

In this thesis, the population pharmacokinetics of telapristone and its active metabolite, CDB-4453 was evaluated using nonlinear mixed effects modeling (NONMEM®). A two-compartment (parent) one compartment (metabolite) mixture model with first order absorption and elimination adequately described the pharmacokinetics of telapristone and CDB-4453. For the Phase I/II pharmacokinetic analysis (effect of renal and hepatic impairment), telapristone was rapidly absorbed with an absorption rate constant (Ka) of 1.26 h-1. Moderate renal impairment resulted in a 74% decrease in Ka. Population estimates for oral clearance (CL/F) for the high and low clearance groups were 11.6 L/h and 3.34 L/h, respectively. Twenty-five percent of the subjects were allocated to the high clearance group. Apparent volume of distribution for the central compartment (V2/F) was 37.4 L, apparent inter-compartmental clearance (Q/F) was 21.9 L/h, and apparent peripheral volume of distribution for the parent (V4/F) was 120 L. The ratio of the fraction of telapristone converted to CDB-4453 to the distribution volume of CDB-4453 (Fmetest) was 0.20/L and apparent clearance of the metabolite (CLM/F) was 2.43 L/h. For the pharmacokinetic analysis evaluating the effect of food; food decreased the Ka of telapristone (Ka for the fed and fasted state was 0.467 and 5.06 h-1, respectively). Population estimates of the high and low CL/F groups were 12.0 L/h and 3.15 L/h, respectively. Thirty-one percent of the subjects were allocated to the high clearance group. V2/F, Q/F and V/4 and Fmetest were 52.8 L, 7.53 L/h, 84.8 L and 0.193/L, respectively. CLM/F was 2.10 L/h. An external validation was performed using the final parameter estimates from the pooled pharmacokinetic analysis (effect of renal and hepatic impairment and the effect of food). From this pharmacokinetic analysis, Ka for the fed and fasted state was 0.299 and 2.35 h-1, respectively. Population estimates for the high and low CL/F groups were 11.6 L/h and 3.22 L/h, respectively. The percentage of subjects allocated to the high clearance group was 29%. V2/F, Q/F, V/4 and Fmetest were 52.8 L, 11.6 L/h and 93.8 L and 0.186/L, respectively. CLM/F was 2.23 L/h. The final model did not meet the requirement for adequate predictability using the external validation dataset. However, the external validation dataset only included samples with limited early time points. Because of the limited sampling times, it is difficult to make a conclusion about the overall adequacy of the model. An external validation dataset with more extensive sampling will be needed in order to better assess the predictability final model. This is the first comprehensive review of the pharmacokinetics of telapristone and CDB-4453.

Background. Clinicians are divided on dosing recommendations when a dose is delayed or missed. For a neuropsychiatric agent like valproic acid (VPA), rational dosing recommendations are of particular importance. VPA is subject to therapeutic monitoring using total concentrations. Due to non-linear binding of VPA to plasma proteins, current dose titration schemes for VPA are empirical. The objectives of this research were to 1- study the effect of missed/delayed doses on steady state concentrations of VPA and 2-design a nomogram that can be used for dose titration based on total VPA concentrations. Methods. 1- A simulation study was conducted to test for different poor compliance scenarios. The effect of missed doses was quantified and used to derive dosing recommendations. 2- A clinical study was carried out in healthy volunteers. Nine volunteers were administered 500, 750 and 1000 mg VPA in a dose escalation study. A nomogram was developed using in vitro plasma protein binding data in all volunteers and tested using dose escalation data. Several delayed/missed doses scenarios were tested in order to validate the simulation model. 3- A revised simulation model was developed using combined information from plasma protein binding and pharmacokinetic analysis of clinical study data. Results and Discussion. Simulation study: Dosing recommendations in the case of a missed or delayed dose are both formulation and dose dependent. Results from the clinical study validated the simulation model and the revised simulation model properly incorporated intra and inter individual variabilities. VPA nomogram: A one-site saturable binding model provided an adequate description of the binding of VPA to albumin. A dosing nomogram for VPA was constructed. To avoid the risk of achieving toxic concentrations, the dose should not be increased by more than 2 fold at a time. The nomogram should be used in conjunction with patient history and clinical response. Conclusions. This research provides dosing recommendations to the clinicians to counsel patients taking preparations of VPA in the event of a missed dose. The use and validation of VPA nomogram will foster the rational use of VPA for the treatment of epilepsy and its role in other neuropsychiatric disorders.

<p>The transport of drugs across the blood-brain barrier (BBB) has been investigated in different species using morphine and morphine-6-glucuronide (M6G) as model compounds. The influence of probenecid on the BBB transport of morphine and M6G was investigated, and the consequences of meningitis and severe brain injury on the concentrations of morphine in the brain were examined. All data were obtained by microdialysis, and data analysis using mathematical models was emphasised.</p><p>Morphine is exposed to active efflux at the BBB in rats, pigs and humans. In addition, the half-life of morphine is longer in the brain than in blood in these species. These interspecies similarities show the predictive potential of the two animal models for the BBB transport of morphine in humans. In the pig the exposure of the brain to morphine was higher in the presence of meningitis than when healthy. This was interpreted as a decrease in the active efflux and an increase in the passive diffusion over the injured BBB. In contrast, there was no significant difference in the concentrations of morphine in the “better” (uninjured) or the “worse” (injured) brain tissue in brain trauma patients. The extent of the transport across the BBB is similar for morphine and M6G. However, co-administration of probenecid only increased the brain concentrations of morphine, demonstrating that morphine and M6G are substrates for different efflux transporters at the BBB. An integrated model for the analysis of data obtained by microdialysis was developed. This model makes fewer assumptions about the recovery, the protein binding and the time of the dialysate observation than a previous model and traditional non-compartmental analysis and should, therefore, yield more reliable parameter estimates.</p><p>Knowledge of the consequences of efflux transporters and disease on the brain concentrations of a drug can be useful for individualising the dosing regimen in patients. </p>

<p>To increase the proportion of patients with successful drug treatment, dose individualisation on the basis of one or several patient characteristics, <i>a priori</i> individualisation, and/or on the basis of feedback observations from the patient following an initial dose, <i>a posteriori</i> individualisation, is an option. Efficient tools in optimising individualised dosing strategies are population models describing pharmacokinetics (PK) and the relation between pharmacokinetics and pharmacodynamics (PK/PD).</p><p>Methods for estimating optimal dosing strategies, with a discrete number of doses, for dose individualisation <i>a priori</i> and <i>a posteriori</i> were developed and explored using simulated data. The methods required definitions of (<i>i</i>) the therapeutic target, <i>i.e. </i>the value of the target variable and a risk function quantifying the seriousness of deviation from the target, (<i>ii</i>) a population PK/PD model relating dose input to the target variable in the patients to be treated, and (<i>iii</i>) distributions of relevant patient factors. Optimal dosing strategies, in terms of dose sizes and individualisation conditions, were estimated by minimising the overall risk. Factors influencing the optimal dosing strategies were identified. Consideration of those will have implications for study design, data collection, population model development and target definition.</p><p>A dosing strategy for <i>a priori</i> individualisation was estimated for NXY-059, a drug under development. Applying the estimated dosing strategy in a clinical study resulted in reasonable agreement between observed and expected outcome, supporting the developed methodology.</p><p>Estimation of a dosing strategy for <i>a posteriori</i> individualisation for oxybutynin, a drug marketed for the treatment of overactive bladder, illustrated the implementation of the method when defining the therapeutic target in terms of utility and responder probability, that is, as a combination of the desired and adverse effects.</p><p>The proposed approach provides an estimate of the maximal benefit expected from individualisation and, if individualisation is considered clinically superior, the optimal conditions for individualisation. The main application for the methods is in drug development where the methods can be generally employed in the establishment of dosing strategies for individualisation with relevant extensions regarding population model complexity and individualisation conditions.</p>

<p>Cefuroxime is a renally eliminated antibiotic used against a variety of different bacterial infections. The pharmacokinetics (PK) for cefuroxime was studied in 97 hospitalized patients using population analysis. To be able to measure cefuroxime in human serum a new sensitive analytical method was developed using mass spectrometry detection. The method was validated and shown to be sensitive and selective. Cystatin C was found to be a better covariate for cefuroxime clearance compared to the traditionally used creatinine clearance (CLcr). This relation might be useful when designing dosing strategies for cefuroxime and other renally eliminated drugs. </p><p>The time-courses of the biomarkers C-reactive protein (CRP), serum amyloid A (SAA), interleukin-6 (IL-6) and body temperature were studied for the first 72 hours of cefuroxime treatment and was related to the duration of illness previous treatment with cefuroxime and to time to step-down of treatment. When duration of illness was short, CRP and SAA were showed increasing levels. None of the biomarkers could be used to differentiate between early or late step-down of therapy.</p><p>By use of known PK and pharmacodynamic (PD) principles, dosing strategies based on CLcr for cefuroxime were estimated using minimization of a risk function. The risk function was constructed with the aim to expose patients to cefuroxime concentration above minimum inhibitory concentration (MIC) for 50 % of the dosing interval and to minimize the amount of drug administered in excess to reach the aim. Based on evaluation using wild type MIC distributions for <i>Escherichia coli</i> and <i>Streptococcus pneumoniae</i> improved dosing strategies were selected.</p><p>In vitro experiments were performed exposing <i>Streptococcus pyogenes</i> to constant concentration of benzylpenicillin, cefuroxime, erythromycin, moxifloxacin or vancomycin. A semi-mechanistic PK/PD model characterizing the time-course of the antibacterial effect was developed using all data simultaneously. Internal validation showed the model being predictive and robust. </p>