QPS ON PK Studies in Patients

Introduction
pkstudy-pic

PK Studies in Patients
There are sometimes distinct subpopulations of people or patients in which drugs behave differently. A key rationale for performing PK studies in patients is to detect these subgroups prospectively.

Examples of such studies are:
  • PK studies in patients with renal insufficiency
  • PK studies in patients with hepatic insufficiency
  • PK studies in critically ill patients
  • PK studies in poor versus extensive metabolizers
  • PK studies in pediatric populations
  • PK studies in elderly populations

PK studies in patients may also be used for the retrospective detection of those distinct subpopulations in which drugs behave differentlythan in the average patient. This end is usually achieved by correlating PK data with other covariate information such as smoking status, age, gender, body weight, renal function, or disease state and/or genetic/genomics data.

Examples of such studies are:
  • PK studies in patients with rheumatoid arthritis
  • PK studies in patients with arrhythmias
  • PK studies in patients with postoperative pain
  • PK studies in patients with type 2 diabetes mellitus

Finally, a third reason to perform PK studies in patients is to test a drug already in the first phases of development, when the adverse event profile of the substance and/or its metabolites carries too serious a safety risk to justify administering it to healthy volunteers. These studies are conducted in patients suffering from the disease the drug is intended to treat.

Examples of such studies are:
  • PK studies of cytostatics in oncology patients
  • PK studies of immunomodulators in cardiac transplant patients
  • PK studies of antipsychotics in schizophrenic patients

Data from all of the above PK studies can either be obtained by extensive blood sampling or sparse sampling. Interpretation of the results of these data will guide prescription decisions and optimal dosage calculations for individual patients.

Bioanalysis

Over the past two decades the introduction of liquid chromatography mass spectrometry has revolutionized drug analysis. This technique enables highly sensitive assays with low limits of quantification, even on very small samples. This advance in bioanalysis has significantly alleviated one of the practical constraints of sampling patient populations: the maximum number of blood samples per patient that can be collected over a given period of time.

PK studies in patients have recently been made even more practical by the introduction of sensitive assays of dried capillary whole blood spots on filter paper. These assays simplify sampling by removing such constraints as: the timing of clinic visits; maximum number of patients that can be recruited in a study; minimum time allowed between consecutive samples for an individual; constrained times for the collection of blood samples; logistical constraints at investigational sites; and cost limitations.

Obviously, bioanalytical requirements will still sometimes mandate particular sampling schedules - for instance, a drug that decomposes in the presence of iron(II) must still be analyzed in plasma, not in dried blood samples. However, the days of suboptimal sampling designs as a consequence of bioanalytical constraints are largely over.

PK Modeling - Optimal Design

PK studies that determine a drug concentration-time profile in a target patient population, using patients who provide only sparse samples at random times, are known as “population PK studies.” Data from all patients is modeled simultaneously using nonlinear mixed effects modeling. This method estimates the mean population PK parameters (e.g. clearance), the inter-individual variability, and the residual variability (which encompasses intra-individual variability, measurement errors and model misspecification).

Although there are now statistical methods for analyzing sparse data from population PK studies, the design of these population PK studies is not straightforward. Many questions arise about how best to conduct these investigations. How many patients should be included? What sampling scheme should be used? How many different sampling schemes are required? How many patients need to be allocated to each sampling scheme, and how should one adjust for covariates that may change over the course of the study? To answer these questions one needs to know the PK profile of the drug, the statistical methods used to analyze the data from population PK studies, and also the practical constraints of sampling the patient population.

Currently, population PK study designs are based on logistical, financial and ethical constraints, as well as prior knowledge of the drug concentration-time profile. These factors are important, but may yet result in a study of “insufficient design” unable to determine the desired PK parameters. Such a study would be unable to resolve the fundamental questions the population PK model was created to answer; the study would be wasted. Thus, there is both a scientific and an ethical imperative to ensure PK studies are designed such that they provide useable results.

References
top
20 years in pharma R&D navigation