Polymorphism can influence every aspect of the properties of a solid including the shelf life, dissolution rate, solubility, formulation properties and processing properties of a solid drug. A deeper understanding of polymorphism and related solid state properties would ensure an improved quality of the materials used throughout drug preparation, dosage form formulation and clinical trials. Therefore, determination of the existence of polymorphs and pseudopolymorphs, characterization of different solid state forms and their respective properties, and controlling the existing form in the resulting formulation all form part of a rapidly growing field within pharmaceutical research and industry. Carbamazepine (CBZ) was the model drug used in this study. FT-Raman spectroscopy was chosen as a main investigative technique in this study to evaluate its potential in monitoring (pseudo)polymorphic conversions in aqueous suspensions in the absence or presence of various pharmaceutical excipients. Partial least squares analysis (PLS) was used for quantitative analysis of the spectral data. Earlier it has been found that CBZ converts rapidly to the dihydrate (DH) when exposed to humidity or water, and this has been reported to be the main reason for the sometimes observed greatly decreased bioavailability of marketed CBZ tablets. In this study, the conversion kinetics of CBZ (forms I, II and III) to DH in aqueous suspension were found to be first order kinetics with an unconverted portion (R� [greater than or equal to] 0.95), where the crystal morphology appeared to play a more important role in its conversion kinetics than the polymorphic form. The influence of pharmaceutical excipients on the conversion of CBZ in aqueous suspension was also explored. For excipients such as methylcellulose (MC), hydroxypropyl methylcellulose (HPMC) and hydroxypropyl cellulose (HPC) which have both a low solubility parameter (< 27.0 MPa[1/2]) and strong hydrogen bonding groups, complete inhibition of the conversion of CBZ was possible even at a very low concentration (0.1 % w/v). Raman spectroscopy showed its high applicability in investigating CBZ conversion kinetics and screening of excipient effects in aqueous environment. It was demonstrated that Raman has a robust nature in quantitative analysis since problems such as different particle size, morphology, and spatial distribution of the two solid state forms of the drug seemed not to have significant influence on Raman scattering. This study has also clarified the relative importance of many contributing factors (type of crystalline form (CBZ or DH), crystal morphology, surface area, and excipient interactions with drug particles) influencing the in vitro dissolution of CBZ. The solid state characterization approach taken in this study will provide a deeper insight into the dissolution performance of drugs and should thus lead to a better understanding of in vitro/in vivo behavior of drugs.

The aims of this Thesis were to investigate early markers of neonatal sepsis and patient-factors affecting the pharmacokinetics and pharmacodynamics (PKPD) of aminoglycosides in the treatment of neonatal sepsis. A prospective cohort study of neonates commenced on gentamicin for suspected sepsis was performed between 1 July 2002 and 28 February 2007. Receiver operator characteristics (ROC) plots were used to assess potential markers of sepsis against culture positive sepsis. When sepsis was first suspected, the most promising tests were interleukin (IL) IL-12(p70) with an area under the curve (95% CI) for the ROC of 0.74 (0.63-0.86), and which (with a cut-off at 75 pg/mL) had a sensitivity (95% CI) of 28% (20-36%) and a specificity of 98% (96-100%). IL-10 had a sensitivity of 17% (10-23%) and a specificity of 99% (97-100%). Retrospective studies of neonates treated with gentamicin, amikacin and netilmicin for suspected sepsis were used to identify patient characteristics that affected aminoglycoside PKPD properties. Population PK modelling used NONMEM� v.5 to determine aminoglycoside clearance (CL) and volume of distribution (V). Logistic regression was used to examine the treatment outcome measures (serum peak and trough concentrations and ototoxicity). Simulations of new dosing regimens were undertaken for netilmicin and amikacin using MATLAB� The final gentamicin PK covariate model gave CL = 0.097 x (current weight/2)[1.3] x (postnatal age/7)[0.29] and V = 1.07 x (current weight/2)[0.8]+ (confirmed sepsis) x 0.13. A 10% increase in gentamicin V in neonates with sepsis was estimated. For amikacin, 17 (35%) of 49 episodes of confirmed sepsis met the treatment failure criteria from 12 (15%) individual patients. The final amikacin PK covariate model was CL = 0.23 x (current weight/2)[0.691] x (postmenstrual age/40)[3.23] and V = 0.957 x (current weight/2)[0.89]. PD analysis determined risk factors linked to hearing impairment in neonates treated with amikacin included: co-medication with vancomycin, high C-reactive protein concentration and low gestational age. Simulation of a new amikacin dosing regimen recommended: 15 mg/kg 36 hourly, 14 mg/kg 24 hourly, and 15 mg/kg 24 hourly, for neonates [less than or equal to] 28 weeks, 29 to 36 weeks, and [greater than or equal to] 37 postmenstrual age, respectively. For netilmicin, the final PK covariate model was CL = 0.192 x (current weight/2)[1.35] x (postmenstrual age/40)[1.03], V = 1.5 x (current weight/2)[0.3]. Simulation of a new optimal dosing regimen for netilmicin was: 5 mg/kg 36 hourly, 5 mg/kg 24 hourly, 6 mg/kg 24 hourly, and 7 mg/kg 24 hourly, for neonates [less than or equal to] 27, 28 to 30, 31 to 33, and [greater than or equal to] 34 weeks postmenstrual age, respectively. IV infusions representing gentamicin administration to neonates of 2.5 kg and 0.5 kg in the NICU setting (30 minutes gentamicin infusion then a 30 minute saline flush) showed the larger neonates received 80% of the drug within 60 minutes. This increased to 90-95% by 75 minutes. However, in extremely low birth weight neonates (0.5 kg), only 60% of the intended gentamicin dose was delivered by 60 minutes (70% by 75 minutes). In conclusion: IL-12(p70) and IL-10 were identified as promising diagnostic tests to confirm sepsis in neonates. Confirmed sepsis caused a 10% increase in V of gentamicin in neonates, suggesting larger initial dosages (mg/kg) are required for effective treatment of neonates with sepsis. Aminoglycoside clearance in neonates is predominantly affected by current weight, postmenstrual age or postnatal age. Adjusting netilmicin and amikacin doses based on current weight, and dosing interval based on both postmenstrual age and current weight improves drug efficacy. Identification of co-medication with vancomycin, low gestational age, and high C-reactive protein during treatment with amikacin increases risk of hearing impairment. The delivery of gentamicin administrated by IV infusion is substantially extended in extremely low birth weight neonates.

Unfractionated heparin (UFH) therapy is frequently used in tertiary paediatric healthcare facilities despite a lack of paediatric-specific research informing the optimal therapeutic intensity, monitoring recommendations or side-effect-profile in infants and children. As a result, the majority of clinical recommendations regarding UFH management in children have been extrapolated from adult evidence. The process of developmental haemostasis, in association with the variable pathogenesis of thromboembolic disease (TED) in children compared to adults, suggests that extrapolation of adult guidelines for UFH management to children is not ideal. / This study hypothesised that the process of developmental haemostasis would influence both the action and effect of UFH in children of different ages. This hypothesis was tested by addressing the following aims: 1. To determine the pharmacokinetics (PK) of UFH in children of different ages; 2. To compare the different methods of monitoring UFH in children of different ages; 3. To identify the impact of competitive plasma binding of UFH in children of different ages; 4. To determine the impact of UFH upon tissue factor pathway inhibitor (TFPI) release in children. / A prospective cohort study of children receiving a single bolus dose of UFH for primary thromboprophylaxis in the setting of cardiac angiography was conducted. Venous blood samples were collected prior to the UFH, then at 15, 30, 45 and 120 minutes post-UFH bolus. Laboratory assays performed included activated partial thromboplastin time (APTT), anti-Xa assay, anti-IIa assay, thrombin clotting time (TCT), protamine titration and TFPI. Levels of two plasma proteins known to competitively bind UFH (vitronectin and platelet factor 4) were determined and the impact of competitive plasma binding upon UFH activity, as measured by the anti-Xa assay, was quantified. A population approach to pharmacokinetic analysis, based on protamine titration results, was performed using WinNonMix™ Professional 2.0.1 (®1998-2000 Pharsight Corporation, Mountain View, CA, USA). Results were analysed according to the following age-groups: less than one year; one to five years; six to ten years; 11-16 years. / Sixty-four children were recruited, ranging in age from six months to fifteen-and-ahalf years. The mean dose/Kg of UFH across the entire cohort was 90.9± 15.5 IU/Kg. / Pharmacokinetic model specifications were systematically assessed, investigating the impact of parameter covariates and different error models upon objective function value and/or curve fitting. A first-order kinetic model best fitted the data. This model used weight 0.75 as the covariate of clearance and total weight as the covariate for volume of distribution. Parameter estimates for clearance and volume of distribution both demonstrated variance from adult and small neonatal PK studies of UFH, however methodological differences in PK analysis techniques limited comparisons. The half-life of UFH reported in this study was consistently and significantly shorter than that previously reported for adults, but longer than that reported for neonates. / All measures of UFH-effect demonstrated a significant and prolonged increase post- UFH bolus. The mean APTT was 261 seconds 102 ± 25 minutes post-UFH, representing a seven-fold increase from the mean baseline APTT (38 seconds). Anti- Xa assay levels were within the therapeutic range for TED management (0.35 to 0.7 IU/mL), or greater, at every post-UFH bolus timepoint. This prolonged UFH-effect was evident to nearly two hours post-UFH bolus, without concurrent UFH infusion. Age-related differences in UFH-response were evident for anti-Xa, anti-IIa and protamine titration results. Furthermore, during periods of high UFH concentration, the ratio of anti-Xa to anti-IIa activity in children less than one year of age significantly favoured UFH-mediated anti-Xa effect over anti-IIa effect (1.9), compared to teenagers (1.3). / This study demonstrated poor correlation between protamine titration and both the anti-Xa assay (r2 = 0.47) and APTT (r2 = 0.56). Use of the anti-Xa assay (0.35 to 0.7 IU/mL) or protamine titration assay (0.2 to 0.4 IU/mL) to establish APTT-based reference ranges for therapeutic management of TED resulted in APTT ranges with upper limits greater than 250 seconds. / No age-related quantitative differences in plasma levels of vitronectin or platelet factor 4 were identified across the childhood years. The addition of dextran sulphate (DS) to ex vivo study samples demonstrated no change in anti-Xa activity in samples collected within 20 minutes of UFH bolus, however a significant increase in anti-Xa activity following the addition of DS was evident at all later timepoints post-UFH bolus. / The measurement of TFPI before and after a single bolus dose of UFH demonstrated children have a similar immediate increase in TFPI activity following intravenous UFH compared to adults. However, the children in this series demonstrated a significantly prolonged level of increased TFPI activity, out to 102 ± 25 minutes post-UFH, compared to that reported in adult patients. / This study has developed the first paediatric-specific PK profile of UFH and has elucidated a number of age-dependent UFH-mechanisms of action that contribute to the previously reported age-dependent response to UFH in children. The results of this study support the hypothesis that developmental haemostasis influences both the action and effect of UFH in children of different ages.

<p>Established quantitative relationships between dose, plasma concentrations and response [pharmacokinetic-pharmacodynamic (PKPD) models] have a high potential in improving therapeutic indices of anticancer drug therapy and in increasing drug development efficiency. PKPD modelling is a helpful tool for characterising and understanding schedule dependence. The aim of this thesis was to develop PKPD models of anticancer drugs for tumour effects and haematological toxicity, which is the most frequent dose-limiting toxicity.</p><p>PK and haematological toxicity after several schedules were studied in rats and semi-physiological PKPD models for the whole time course of myelosuppression were developed from animal and patient data. The possibility to implant hollow fibres filled with tumour cells in immunocompetent rats was investigated for simultaneous assessment of PK, tumour response and haematological toxicity. Population data analyses were performed using the software NONMEM. </p><p>When all injections were administered within eight hours, fractionated schedules of 5-fluorouracil and epirubicin produced similar haematological toxicity in rats as a single dose, when the non-linear PK of 5-fluorouracil was accounted for. When the time interval was extended to two days for 5-fluorouracil, the fractionated regimens were more toxic. </p><p>The developed semi-physiological PKPD models included transit compartments that mimic maturation stages in bone marrow and explain the time lag. Feedback mechanisms characterised the rebound. The models successfully described myelosuppression in patients (DMDC) and rats (5-fluorouracil), after different administration schedules. Further developments made it possible to characterise the time course of myelosuppression after administration of each one of six different drugs, with parameters related to the haematopoietic system consistent across drugs. </p><p>The developed hollow fibre model in immunocompetent rats was successfully applied to monitor PK, toxicity and the time course of antitumour effects. PKPD modelling illustrated that the schedule dependence of the anticancer agent CHS 828 is partly due to dose-dependent bioavailability and partly due to a schedule-dependent PD effect.</p>

<p>Acid related diseases like GERD, duodenal-and gastric ulcers and H. Pylori-positive peptic ulcer disease are primarily managed by reducing gastric acidity. Irreversible proton pump inhibitors (PPIs) inhibit gastric acid secretion effectively throughout the day by irreversibly inhibiting the gastric proton pump, H+, K+-ATPase, in the parietal cells. Reversible gastric proton pump inhibitors are under development, but have not yet reached clinical use.</p><p>The pharmacokinetic/pharmacodynamic (PK/PD) relationships of these compounds are nonlinear, with a delay in the effect-time profile compared to the plasma concentration-time course. PK/PD-modelling was used to characterize and quantify the pharmacological effect with regard to onset, intensity and duration of effect. Models based on functional data, that discriminate between drug-and system-specific parameters, were developed. </p><p>In general, the plasma concentration-time course for each individual was approximated by linear interpolation between time-points and served as input into the pharmacodynamic models. A turnover model of irreversible inhibition of gastric acid secretion by omeprazole in the dog described the data well. The model was challenged and found to be robust under different experimental conditions. This model could predict the effect following different exposure of omeprazole and following different histamine provocation. Different fitting approaches (naïve pooling, standard two-stage and nonlinear mixed effects modelling) were compared and resulted in similar parameter estimates. For the reversible inhibitors, a kinetic binding model was finally selected. With a binding model the delay in the effect-time profile is explained by prolonged binding to the enzyme. </p><p>Use of these results in drug development can be helpful with regard to selection of drugs for further development and to predict the first clinical dose.</p>

This dissertation describes formulation of a gastric retention device (GRD) and sustained release (SR) hydrochlorothiazide beads at Oregon State University. Formulation condition and amounts of excipients had significant influence on characteristic of the GRD. The GRD containing SR hydrochlorothiazide beads was employed to assess bioavailability/bioequivalency study in healthy subjects. An original GRD was retained in the stomach with food and completed the drug release. However, this original GRD failed to stay on an empty stomach, leading to lower bioavailability than an immediate release (IR) tablet due to insufficient rigidity. The original device was modified to be more rigid, and this more rigid device successfully stayed on an empty stomach and achieved completion of the drug release at a slow release rate, the bioavailability and the drug effect on diuresis increased compared to the drug in an IR tablet. Less amount of hydrochlorothiazide at a slow release rate achieved equivalent diuresis to higher amount of the drug at a rapid release rate, which indicated slow drug release resulted in higher efficiency of hydrochlorothiazide. In vivo/in vitro correlation of hydrochlorothiazide in a modified GRD and an IR tablet was established to predict in vivo profile with in vitro dissolution profile prior to a clinical study. Pharmacokinetic/pharmacodynamic of nicotine was reviewed in terms of a relationship between plasma nicotine concentrations and pharmacological changes including heart rate and craving. Considering craving and development of tolerance to nicotine effect on cardio-acceleration, a dosing regimen with a combination of rapid input and constant slow input was suggested to improve smoking cessation. A simulation study was carried out to verify the current regulatory policy on assessing bioequivalency of enantiomeric drugs. First-order dissolution and absorption process, and nonlinear stereo-specific pre-systemic and systemic metabolism was taken into account to establish a pharmacokinetic model for the simulation. Four different dissolution profiles, within- and between-subject variability, dose and sample size were considered to simulate 1000 cross-over bioequivalency trials under standard bioequivalency criteria. Probability of false positives was determined to evaluate the current policy. The simulation study validated the importance of individual enantiomer pharmacokinetic for assessing bioequivalency study of the chiral drugs. / Graduation date: 2003

Graduation date: 1993

Dissertation (Ph.D.)--University of Stellenbosch, 1996. / Bibliography. Also available via the Internet.

Established quantitative relationships between dose, plasma concentrations and response [pharmacokinetic-pharmacodynamic (PKPD) models] have a high potential in improving therapeutic indices of anticancer drug therapy and in increasing drug development efficiency. PKPD modelling is a helpful tool for characterising and understanding schedule dependence. The aim of this thesis was to develop PKPD models of anticancer drugs for tumour effects and haematological toxicity, which is the most frequent dose-limiting toxicity. PK and haematological toxicity after several schedules were studied in rats and semi-physiological PKPD models for the whole time course of myelosuppression were developed from animal and patient data. The possibility to implant hollow fibres filled with tumour cells in immunocompetent rats was investigated for simultaneous assessment of PK, tumour response and haematological toxicity. Population data analyses were performed using the software NONMEM. When all injections were administered within eight hours, fractionated schedules of 5-fluorouracil and epirubicin produced similar haematological toxicity in rats as a single dose, when the non-linear PK of 5-fluorouracil was accounted for. When the time interval was extended to two days for 5-fluorouracil, the fractionated regimens were more toxic. The developed semi-physiological PKPD models included transit compartments that mimic maturation stages in bone marrow and explain the time lag. Feedback mechanisms characterised the rebound. The models successfully described myelosuppression in patients (DMDC) and rats (5-fluorouracil), after different administration schedules. Further developments made it possible to characterise the time course of myelosuppression after administration of each one of six different drugs, with parameters related to the haematopoietic system consistent across drugs. The developed hollow fibre model in immunocompetent rats was successfully applied to monitor PK, toxicity and the time course of antitumour effects. PKPD modelling illustrated that the schedule dependence of the anticancer agent CHS 828 is partly due to dose-dependent bioavailability and partly due to a schedule-dependent PD effect.

Acid related diseases like GERD, duodenal-and gastric ulcers and H. Pylori-positive peptic ulcer disease are primarily managed by reducing gastric acidity. Irreversible proton pump inhibitors (PPIs) inhibit gastric acid secretion effectively throughout the day by irreversibly inhibiting the gastric proton pump, H+, K+-ATPase, in the parietal cells. Reversible gastric proton pump inhibitors are under development, but have not yet reached clinical use. The pharmacokinetic/pharmacodynamic (PK/PD) relationships of these compounds are nonlinear, with a delay in the effect-time profile compared to the plasma concentration-time course. PK/PD-modelling was used to characterize and quantify the pharmacological effect with regard to onset, intensity and duration of effect. Models based on functional data, that discriminate between drug-and system-specific parameters, were developed. In general, the plasma concentration-time course for each individual was approximated by linear interpolation between time-points and served as input into the pharmacodynamic models. A turnover model of irreversible inhibition of gastric acid secretion by omeprazole in the dog described the data well. The model was challenged and found to be robust under different experimental conditions. This model could predict the effect following different exposure of omeprazole and following different histamine provocation. Different fitting approaches (naïve pooling, standard two-stage and nonlinear mixed effects modelling) were compared and resulted in similar parameter estimates. For the reversible inhibitors, a kinetic binding model was finally selected. With a binding model the delay in the effect-time profile is explained by prolonged binding to the enzyme. Use of these results in drug development can be helpful with regard to selection of drugs for further development and to predict the first clinical dose.