Assessment of methods to identify sources of interindividual pharmacokinetic variations.

The advantages and limitations of the 2 most commonly used methods to investigate interindividual pharmacokinetic variations are reviewed. The first method is based on pharmacokinetic comparisons made after repeated administration of a model drug such as antipyrine, before, during and after imposition of a carefully controlled environmental perturbation. A principal virtue of the test is the use of each subject as a control. Subjects are usually under near basal conditions with respect to factors capable of altering hepatic drug-metabolising capacity. Exceedingly sensitive, the test yields highly reproducible results. It has been useful as a research tool in identifying environmental factors for which dose-response curves can be generated and compared. However, the test requires careful selection and control of subjects, and it may be hazardous to extrapolate results to subjects under different, non-basal, environmental conditions. This method most frequently involves antipyrine as the test compound, but other drugs can and have been used. The results disclose that many host factors that influence antipyrine disposition also affect the disposition of other drugs metabolised by hepatic mixed-function oxidases. Recent refinement of the antipyrine test involves measurement of the rate constant for formation of each of the 3 main metabolites of antipyrine. Sensitivity and specificity of the test are increased through examination of the effect of each factor on a separate hepatic cytochrome P-450. Due to the labouriousness of this procedure and its requirement for several days of urine collection from each subject, metabolite analysis will probably remain an experimental method not applicable for screening populations. The second method involves a particular model based on multiple regression analysis. Relying on correlations with historical data of a qualitative nature, previous applications of this method have been retrospective, rather than prospective. Several such correlations could not be confirmed in normal subjects under the conditions of a controlled prospective experiment. Thus, prospective studies need to be performed to check results obtained with this method. The model used appears to enjoy certain advantages, including speed, simplicity, and ease of execution.(ABSTRACT TRUNCATED AT 400 WORDS)