Therapeutic drug monitoring databases for postmarketing surveillance of drug-drug interactions.

Drug-drug interactions can be associated with patient morbidity due to either increased toxicity or a potentially ineffective concentration. Because interactions cannot always be anticipated during drug development and actual patients receiving a drug for therapeutic use often differ from those included in clinical trials, postmarketing surveillance is essential. Therapeutic drug monitoring (TDM) databases offer a unique opportunity in this respect. Prerequisites for TDM databases to provide valid information in a pharmacoepidemiological perspective include the following: precise description of exposure to the potentially interacting drugs; measurement of parent compound and active metabolites through accurate and precise analytical techniques; documentation of relevant patient characteristics that may act as confounding factors (e.g. gender, age, smoking habits); repeated assessments over time if possible; and sound pharmacokinetic framework for data selection, analysis and interpretation. The contribution of TDM to the documentation of drug-drug interactions takes advantage of different possible study designs, discussed on the basis of recently published studies. The single case report plays an important role as an alert signal. It is illustrated for a patient on long-term treatment, who displayed an unexpectedly high clozapine concentration after the introduction of ciprofloxacin comedication. The prospective on and off comedication panel study shows advantages in terms of carefully selected inclusion criteria and control of treatment modalities. A study of the thioridazine-fluvoxamine interaction is presented, with patients followed on thioridazine monotherapy, after introduction of fluvoxamine and after its discontinuation. The main advantage of the retrospective large-scale TDM database screen is representativeness of patients actually treated, whereas drawbacks are related to quality of data and suitability for valid interpretation. Such an approach is illustrated by a review of data collected over 10 years of routine TDM that allowed documenting induction of nortriptyline metabolism by carbamazepine and inhibition by several phenothiazines. Finally, population pharmacokinetics is well suited to observational data collected for TDM purpose, provided quality is ascertained. Focus is placed on interindividual variability and relationship between pharmacokinetic parameters and patient characteristics, including comedication. The population approach is discussed with respect to a study that documented a 32% increase of haloperidol clearance associated with anticonvulsant comedication, in addition to effects of age and bodyweight. Among factors to consider for improved effectiveness in the use of TDM databases for postmarketing surveillance of drug-drug interactions, integration of efficacy and safety data in future studies and communication of expert recommendations to prescribing physicians are essential.

Population pharmacokinetics of clomipramine, desmethylclomipramine, and hydroxylated metabolites in patients with depression receiving chronic treatment: model evaluation.

Because metabolites play a major role in the clinical response to clomipramine, the objective of the current study was to develop a population model and evaluate its performance to describe the pharmacokinetic profiles of clomipramine (C) and its active metabolites desmethylclomipramine (DC), 8-hydroxy-clomipramine (OHC) and 8-hydroxy-desmethylclomipramine (OHDC). A first sample of 14 patients served for development of a 2-molecule C and DC model, which was shown to provide reasonable estimates of AUC-based clearances, as well as precise estimation of interindividual variability. Simulated data, generated to mimic a semi-rich sampling design and chronic treatment with clomipramine, indicated that clearance estimation was feasible under routine treatment conditions. A second sample of 30 patients, recruited prospectively and followed for a median 4-week period, was used to extend the 2-molecule model to a 4-molecule model. Goodness-of-fit assessment revealed that model-predicted concentrations were reasonably close to observed concentrations for a majority of patients. Interindividual variability was 50% to 60% for hydroxylation and desmethylation clearances, and residual variability was 30%. The proposed model incorporates much of what is known about the metabolism of clomipramine and may valuably integrate the influence of genetic and environmental factors on each metabolic pathway.

Modelling during drug development.

With the advancement of both biological and computer sciences, new drug development faces the challenge to integrate a huge amount of knowledge accumulated from the very early quantitative structure-activity relationship investigations of the candidate molecule to the large scale clinical trials in patients. Whereas pharmacokinetics and pharmacodynamics are fields in which modelling has long demonstrated its value, its potential in many other areas of drug development has recently been the object of intensive scientific activity. The present review places emphasis on these newer applications; it includes the opinion of many experts in often highly specialised areas such as in vitro to in vivo extrapolation, toxicokinetics, non-continuous response models, population approaches and computer assisted simulation of clinical trials. It is most probable that in the near future many of these areas of research will be the objects of intensive and interesting developments. This will undoubtedly lead to improve developmental strategies for new drugs as well as more individualised pharmacological strategies for patients.

Cultural differences: implications on drug therapy and global drug development.

INTRODUCTORY REMARKS: Discussing the "inter-ethnicity" of kinetics and actions of drugs is fraught with terminology problems. It is, however, generally accepted that "ethnicity" covers the influences of factors genetically and culturally transmitted. The study of inter-ethnic variability in drug response addresses the problem of distinguishing variability factors that are common to one particular group of individuals from those which are not specifically shared. DIFFERENCES IN DAILY DOSES BETWEEN DIFFERENT GEOGRAPHIC REGIONS: It is well known that for a number of drugs, the daily dose prescribed in Japan is lower than in the US and Europe. Presently, independent surveys strongly indicate that for a majority of drugs dose differences are not the result of pharmacokinetic differences. In addition, they indicate that inter-ethnic differences do not seem to be larger than intra-ethnic variations. The differences observed for daily doses must thus be found elsewhere than in pharmacogenetic traits. DIFFERENCES IN DIAGNOSES: The most important impediments encountered in the evaluation of minority patients include differences with respect to language, communication style, cultural belief. The same problems arise if studies performed in different geographic areas are compared, socio-economic aspects play then an even greater role. Language problems arise differently if minorities are evaluated and compared to a majority of patients living in the same country or if clinical studies are performed in different regions. Communication styles also differ markedly between cultures. As an example, certain gestures may be considered as disrespectful or insulting by some ethnic groups and constitute normal behavior in others. RATING SCALES: Ethnicity clearly plays a role on the cross-cultural use of rating scales. Sophisticated rating scales established and validated in Western culture must undergo culturally sensitive revision and rigorous evaluation before their use in non-Western culture. EFFICACY-SAFETY ASSESSMENT: As an example, the assessment of risk and benefit is different in Japan, Europe and the United States. In Japan, safety is given a greater weighting relative to efficacy than in the two other regions. PLACEBO/NOCEBO EFFECTS: Placebo and nocebo effects are difficult to study, even in the absence of any cultural difference. They are even more so if ethnicity is concerned. PATIENT COMPLIANCE: Clinicians treating cross-cultural patients must carefully explore the beliefs held by their patient regarding illness causality and treatment expectations. CONCLUDING REMARKS: There are many unanswered questions in the field of inter-ethnic variability in drug response. The present overview will not pretend to have given specific answers, but it is hoped that it will point to some areas where more research is needed, in particular in the area of methodologies to take inter-ethnicity into account during drug development.

Enantioselective analysis of methadone in saliva by liquid chromatography-mass spectrometry.

Saliva was tested and evaluated as a biological matrix for methadone (Mtd) monitoring. Conventional method using a narrow bore C18 column, and an enantioselective method using a narrow bore alpha1-acid glycoprotein column, were developed using liquid chromatography coupled with a mass spectromeric (MS) detector. After optimisation of MS conditions by flow injection analysis, selected ion monitoring detection was used to enhance sensitivity. The total Mtd concentration and the enantiomeric ratio in saliva were validated using an experimental design. The methods were applied to samples provided by heroin addicts undergoing a Mtd treatment. Results on total Mtd determination showed a very poor correlation between saliva and serum, whereas the enantiomeric ratios of Mtd gave a very good one.

Development of validated stereoselective methods for methadone determination in clinical samples.

A stereoselective analysis of methadone (Mtd) in whole blood and serum was developed using liquid chromatography on a protein based chiral stationary phase. Liquid-liquid extraction (LLE) and solid phase extraction methods were applied before chromatographic analysis. The extraction procedure, as well as the choice of the biological matrix, showed significant differences in the extraction yield and in the precision of the assays. Serum was selected for this assay and LLE was chosen as the preparation step because of its simplicity and rapidity. The total procedure was validated and applied to clinical samples. Samples taken from 45 heroin-addicted patients were analyzed. A correlation was found between the dose administered and Mtd concentration (total and R-form), but interindividual variability of the total normalized Mtd was seen (concentration varied from 90 to 530 ng/ml). Furthermore, two populations were apparently observed with a mean Mtd concentration of 200 and 475 ng/ml, respectively. Stereoselective analyses showed that more than 50% of the patients presented a nonracemic ratio, and particularly about 25% showed a preferential metabolism of the active R-Mtd enantiomer. Therefore, the stereoselective determination of Mtd is necessary to improve the quality of the treatment of heroin addiction.

Clomipramine concentration as a predictor of delayed response: a naturalistic study.

OBJECTIVES: The aim of the present study was to characterise onset of response to clomipramine treatment in a naturalistic setting and to investigate the relationship between concentration and delayed response, postulated to reflect drug-specific response to antidepressant therapy. METHODS: Ninety-eight depressed patients were prescribed clomipramine in an open flexible manner and followed for depressive symptoms (Montgomery-Asberg depression scale) over a maximum 12 weeks follow-up period. All patients had at least one concentration measurement for therapeutic drug monitoring purpose. RESULTS: Firstly, survival analysis revealed a probability of 15.4% for patients not to show 50% improvement over baseline by week 12, and thus to be considered as non-responders. Median time to onset of response was 31 days for responders, indicating a relatively high probability of delayed response under routine treatment. Secondly, pattern analysis indicated a significant association between early and abrupt response on the one hand and delayed and gradual response on the other. A tendency towards an association between delayed and persistent response was also observed. Finally, receiver operating characteristics analysis allowed identification of a highly significant lower threshold of 230 ng x ml(-1) for the sum of clomipramine and desmethyl-clomipramine, as measured at week 3, with respect to response from week 3 onward. Predictive values were 68.8% and 81.0% for concentrations above and below this threshold to predict delayed response and non-response, respectively. Thresholds were 55 ng x ml(-1) for parent compound and 180 ng x ml(-1) for metabolite. CONCLUSION: This approach supports the hypothesis that delayed response may be concentration dependent and thus may reflect true drug effect. As a consequence, monitoring clomipramine concentration about 3 weeks after initiation of therapy may valuably contribute to help clinicians decide about the adequacy of ongoing therapy.

Therapeutic drug monitoring of risperidone using a new, rapid HPLC method: reappraisal of interindividual variability factors.

Because of the enormous gap between premarketing studies in physically healthy subjects and clinical practice in patients, the present study reconsidered interindividual variability factors affecting risperidone concentrations under routine therapeutic drug monitoring conditions. The study included 92 patients, 27% of whom were 70 years or older. The patients received risperidone orally (dose range, 0.5-11 mg per day) and had concentrations of risperidone and the active metabolite 9-hydroxyrisperidone measured at steady state by a new, rapid, and sensitive method of high-performance liquid chromatography (HPLC). After normalization to a dose of 4 mg/day, median concentrations were 2.9 ng/ml (80% range, 0.9-27.9 ng/ml) for the parent compound and 24.1 ng/ml (80% range, 12.0-57.6 ng/ml) for the metabolite. When considering linear regression models, age was identified as a major source of interindividual variability, with expected increases of 340% and 220% for concentrations of parent compound and metabolite, with age increasing from 20 to 80 years. Body weight provided an additional significant contribution to the variability of 9-hydroxyrisperidone concentration, a 20-kg higher body weight associated with a concentration decrease of 23%. Serotonin-specific reuptake inhibitor (SSRI) comedication (fluoxetine, two patients; citalopram, two patients; paroxetine, one patient; fluvoxamine, one patient) was significantly associated with 4.6-fold higher concentrations of parent compound, in keeping with an inhibitory action on CYP2D6 enzyme. Significantly higher concentrations of 9-hydroxy-risperidone (+ 29%) were also observed in the 17 patients with biperiden comedication. Therapeutic drug monitoring data, collected in patients representative of the population for which the drug was intended, allowed us to quantify the dose reduction needed in elderly patients and thus provided valuable information in addition to the one collected during premarketing studies performed with strict inclusion and exclusion criteria.

Optimized release of dexamethasone and gentamicin from a soluble ocular insert for the treatment of external ophthalmic infections.

In the case of external ophthalmic infections, repeated instillations of antibiotics are required to reach therapeutic level, above the minimal inhibitory concentration (MIC). An additional administration of a corticosteroid is often needed, in order to limit the precorneal damages caused by the infection. However, repeated administration of a corticosteroid can increase intraocular pressure and thus lead to glaucoma. To overcome the disadvantages of separated and repeated instillations of two products and to avoid the side effects of dexamethasone, a soluble insert containing gentamicin sulfate and dexamethasone phosphate was developed. The new system ensures the concomitant release of the two drugs during the first 10 h of treatment, followed by an adequate concentration of gentamicin sulfate, above the MIC of 4.0 microgram ml-1, during 50 h, due to a combination of gentamicin sulfate with cellulose acetate phthalate, which reduces the solubility of gentamicin.

Comparative pharmacokinetic study of a floating multiple-unit capsule, a high-density multiple-unit capsule and an immediate-release tablet containing 25 mg atenolol.

The purpose of this study was to evaluate the possible advantages of floating and high-density dosage forms and their influence on pharmacokinetic parameters. Atenolol was chosen as a model drug because of its poor absorption in the lower gastrointestinal tract. Three formulations containing 25 mg atenolol, a floating multiple-unit capsule, a high-density multiple-unit capsule, and an immediate-release tablet were compared with respect to estimated pharmacokinetic parameters. The two multiple-unit dosage forms were composed of compressed minitablets and had sustained release properties. The bioavailability of the two gastroretentive preparations with sustained release characteristics was significantly decreased when compared to the immediate-release tablet. The floating minitablets seemed to be retained longer in the stomach than the high-density dosage form. The first atenolol concentration detectable in the plasma and the time to peak Tmax were delayed for the floating dosage form. For the parameters Cmax and AUC 0-infinity, the lower limit of the 90% confidence interval was outside the bioequivalence range (0.80-1.25). This study showed that it was not possible to increase the bioavailability of a poorly absorbed drug such as atenolol using gastroretentive formulations. Atenolol absorption was delayed and the maximum plasma concentration was diminished.

Density of guanine nucleotide-binding proteins in platelets of patients with major depression: increased abundance of the G alpha i2 subunit and down-regulation by antidepressant drug treatment.

The aim of this study was to quantitate the density of guanine nucleotide-binding (G) protein subunits (inhibitory G alpha i, stimulatory G alpha s, G alpha q/11, and G beta) in platelets of unipolar depressed patients to assess the status of these signal transduction proteins in depression and the effects of antidepressant drug treatment. Blood platelets were collected from 22 drug-free depressed patients and 22 age- and sex-matched healthy controls. The levels of the various G protein subunits were assessed by immunoblotting techniques. The immunoreactivity of G alpha 12 was increased (41%) and that of G alpha i3 decreased (25%) in platelets of depressed patients. The levels of other G protein subunits (G alpha s, G alpha q/11, G beta) did not change significantly with respect to those of control subjects. Chronic administration of cyclic antidepressant drugs (citalopram, clomipramine, imipramine) decreased the immunoreactivity of the up-regulated G alpha i2 protein (31%). Since platelet G alpha i2 is in line with the existence of supersensitivity of these receptors in major depression.

Kinetics of blood component adsorption on poly(D,L-lactic acid) nanoparticles: evidence of complement C3 component involvement.

After intravenous administration, nanoparticles suffer a major drawback in that they are rapidly and massively taken up by the cells of the mononuclear phagocyte system. The mechanisms involved in the opsonization, adhesion, and internalization of biodegradable nanoparticles by the mononuclear phagocyte system are still poorly understood. In this work, the kinetics of blood protein adsorption onto nanoparticles of poly(D,L-lactic acid) prepared by the salting-out technique was investigated. Nanoparticles of 312 nm were incubated for variable periods of time (5-60 min) in human serum and citrated plasma. After incubation, the particles were washed and the proteins detached from them, denatured, and analyzed by two-dimensional polyacrylamide gel electrophoresis. In plasma, the predominant protein was immunoglobulin G (IgG), and the amount adsorbed was not dependent on incubation time. Albumin amounts were high for short incubation periods but decreased as a function of time, whereas apolipoprotein E levels increased significantly as a function of the incubation period. Owing to the possible complement cascade inactivation by addition of citrate to plasma, the kinetics of adsorption was also evaluated in serum. In this medium, adsorption of complement C3 components onto the surface of the nanoparticles was clearly evidenced by spots of increasing intensity and area, reaching levels comparable to those of the omnipresent IgG. This result confirms the important role of complement components in the opsonization process of poly(D,L-lactic acid) particles.

Therapeutic drug monitoring databases for postmarketing surveillance of drug-drug interactions: evaluation of a paired approach for psychotropic medication.

The present study investigated the potential of therapeutic drug monitoring data to document pharmacokinetic drug interactions with psychotropic medication, both in terms of methodology and applicability. It focused on 105 patients exposed to one of five agents known for their capacity to induce (phenytoine, phenobarbital, and carbamazepine) or to inhibit (thioridazine and levomepromazine) the metabolism of psychotropic drugs. These patients were matched by gender, age, and monitored psychotropic medication to 105 patients randomly selected from a pool of subjects nonexposed to target comedication. Such a paired approach was shown to be effective in reducing variability for a majority of substances. Power analysis suggested that eight to 10 pairs of exposed and nonexposed patients would effectively allow the detection of twofold effects of interacting substances. In keeping with the literature, analysis of the ratios of dose-normalized exposed to nonexposed concentrations indicated that phenothiazine comedication led to significantly higher concentrations of desmethylated metabolites but not parent compounds, when clomipramine, imipramine, or amitriptyline were administered. A similar, as yet undocumented interaction was observed for the tetracyclic antidepressant mianserine. In contrast, the present study revealed that maprotiline concentrations were increased, but its metabolite was largely unaffected by phenothiazine comedication. Increased concentrations were also observed for moclobemide, but not citalopram or its metabolite. In addition to its long recognized capacity to decrease haloperidol concentrations, carbamazepine was shown to induce the metabolism of clopenthixol and possibly flupenthixol. The consistency of such a picture substantiates the need to consider therapeutic drug monitoring databases as cost-effective and reliable tools for postmarketing surveillance.

Drug metabolism as a confounding factor in PK/PD population approaches.

Pharmacokinetic/pharmacodynamic population approaches aim at establishing relationships between dose, drug concentration of active principles and clinical response, accounting for factors responsible for inter-individual variability. Additional difficulties in the presence of metabolites include the need to decide a priori which metabolites should be monitored according to their respective role in efficacy and/or toxicity, the need to select appropriate selective analytical methods, and the requirement for more complex pharmacokinetic and/or pharmacodynamic models. These points are discussed more extensively for psychotropic drugs, and in particular for the active metabolites of phenothiazines, for reduced haloperidol, for the desmethylated and hydroxylated metabolites of tricyclic antidepressants, and for the desmethyl metabolites of the serotonin-specific reuptake inhibitors fluoxetine and citalopram.

Therapeutic drug monitoring and drug-drug interactions: a pharmacoepidemiological perspective.

The present study investigates the potential of therapeutic drug monitoring databases to document co-medication as a possible risk factor for subtherapeutic or excessively high concentrations of psychotropic drugs. Exposure was defined with respect to co-medication including one of five agents known for their capacity to induce (phenytoin, phenobarbital and carbamazepine) or to inhibit (thioridazine and levomepromazine) the metabolism of psychotropic drugs. 87 patients exposed to such co-medication were matched by sex, age and monitored psychotropic medication with 87 patients randomly selected from a pool of subjects whose co-medication did not include any substance known to interact. Outcome was defined with respect to dose-normalized concentrations being below or above therapeutic range. When taking all psychotropic drugs together, the estimated relative risk to reach concentrations above the therapeutic range was 7.8 for patients exposed to phenothiazine co-medication. The relative risk to remain at subtherapeutic level was 2.7 for patients with inducers. When considering the different psychotropic drugs separately, a coherent picture was observed, with increased risk ratios for all substances.

Investigation of xenobiotic metabolism by CYP2D6 and CYP2C19: importance of enantioselective analytical methods.

Investigations into the genetic polymorphism of drug metabolism have involved specific models to screen poor and extensive metabolisers of xenobiotics. Debrisoquine, sparteine, S-mephenytoin and dextromethorphan are particularly well known. They have been extensively described in the literature and are used to phenotype human subjects before performing investigations with new drugs which are believed to be under the control of a genetic polymorphism. Dextromethorphan, debrisoquine and sparteine are good substrates for CYP2D6, whereas the S-enantiomer of mephenytoin is a good substrate for CYP2C19, both being two isozymes of cytochrome P-450. In many drugs, the hepatic microsomal oxidative metabolism involving stereogenic centres congregates either with CYP2D6 or with CYP2C19 or, in certain cases, with both of them. The availability of both CYP2D6 from poor and extensive metabolisers and an enantioselective assay would allow genetic polymorphism in drug biotransformation to be investigated in vitro ex vivo at an early stage of drug development before the IND (investigational new drug). Single-dose investigations in vivo can also be performed when only minimal pre-clinical toxicological data are available and produce more reliable results than in vitro studies. This paper focuses on the problem of genetic polymorphism in drug development and specifically discusses some relevant knowledge gained in the last two decades on enantioselective bioassays. Specific examples are given.

New alterations of serum glycoproteins in alcoholic and cirrhotic patients revealed by high resolution two-dimensional gel electrophoresis.

Plasma protein are synthesized and secreted by the liver. Several reports have shown that excessive consumption of ethanol interferes with the hepatic protein synthesis and/or secretion. This study was undertaken to identify the plasma/serum proteins altered in two groups of patients with different alcohol-related diseases: actively drinking alcoholic patients group without liver disease and alcohol cirrhotic patients group. Two-dimensional gel electrophoresis was used to separate proteins with high resolution. Proteins were detected by silver staining and glycoproteins were specifically visualized and analyzed after lectin blotting followed by chemiluminescence detection. Different protein alterations were identified in each group of patients. In the alcoholic group, two new glycosylation modifications of serum proteins were identified. An abnormal microheterogeneity of haptoglobin and alpha1-antitrypsin was detected in the serum of all alcoholic patients. We also characterized by two-dimensional gel electrophoresis the carbohydrate deficient transferrin. The modifications of haptoglobin, alpha1-antitrypsin and transferrin present a similar change of charge and molecular weight in the two-dimensional gel electrophoresis pattern. These qualitative estimations support the hypothesis of a general mechanism of liver glycosylation alteration of serum proteins induced by excessive alcohol consumption. The immunoglobulin alterations were easily visualized and identified for the cirrhotic and the alcoholic patients. And finally, the decrease of haptoglobin and albumin spots for cirrhotic patients was confirmed.

The use of population pharmacokinetics in drug development.

Currently, there is an increasing focus on the implementation of pharmacokinetic-pharmacodynamic (PK-PD) studies and modelling as essential tools for drug development. Strategies involving specifically the population approach, which are based on relatively recent statistical methodology (e.g. nonlinear mixed effects modelling, NONMEM) have been advocated for investigating pharmacokinetic and pharmacodynamic variability as well as dose-concentration-effect relationships. The present article outlines this approach, and discusses how it can be implemented within the framework of the studies currently performed as part of the clinical phases of new drug development. It also considers study design and performance, based on real-life experiences. Population approaches, if designed carefully and early, as part of the planning of the drug development programme, are expected to play a significant role at every phase of the programme and to contribute to providing information that is valuable for registration purposes. Statistical methodology and software are now widely available. However, practical issues such as integration of the population approach within existing protocols, quality control of the data, timing of laboratory and statistical analyses, as well as resource allocation, remain legitimate concerns to be considered in prospective studies.

Antidepressants and drug-metabolizing enzymes--expert group report.

Antidepressant drugs are extensively metabolized. Consequently, the biotransformation pattern of antidepressants has an important influence on their clinical properties, i.e., pharmacokinetics, toxicity, drug-drug interactions, side-effect profile and last but not least therapeutic efficacy. It was against this background that a multidisciplinary group of experts discussed the clinical relevance of the rapidly increasing body of knowledge of antidepressant-metabolizing enzymes. The variability of the response of a given individual to an antidepressant is determined genetically and by the environment. Genetic polymorphism of drug-metabolizing enzymes and inhibition by other substrates may affect the enzymatic biotransformation of antidepressants. In vitro assay techniques allow an estimation of the potential variability in clinical response to antidepressants and a reasonable prediction of the drug-drug interaction patterns. The results of in vitro tests should therefore be considered early in the development of an antidepressant as a background for designing clinical studies (treatment schedules and dosing). Physicians should have an understanding of the relevance of genetic polymorphism for clinical practice. Education is needed in order to fill the existing gaps in knowledge about antidepressant-enzyme interactions and their application in daily treatment practice. The information on potential drug interactions determined by genetic polymorphism and based on studies with enzymes should be increasingly contained in drug compendia.

Practical experience and issues in designing and performing population pharmacokinetic/pharmacodynamic studies.

An expert meeting to discuss issues relating to the design of population pharmacokinetic/pharmacodynamic (PK/PD) studies was held in Brussels in March 1995, under the auspices of the European Co-operation in Science and Technology (COST), Medicine (B1) programme. The purpose of the meeting was to discuss the experts' experience in designing and performing population PK/PD studies. The topics discussed were current practice, logistical issues, ensuring the accuracy of data, covariate assessment, communication, and protocol design.