Inter- and intraindividual variability in pharmacokinetics of most drugs is largely determined by variable liver function as described by parameters of hepatic blood flow and metabolic capacity. These parameters may be altered as a result of disease affecting the liver, genetic differences in metabolising enzymes, and various types of drug interactions, including enzyme induction, enzyme inhibition or down-regulation. With the now known large number of drug metabolising enzymes, their differential substrate specificity, and their differential induction or inhibition, each test substance of liver function should be used as a probe for its specific metabolising enzyme. Thus, the concept of model test-substances providing general information about liver function has severe limitations. To test the metabolic activity of several enzymes, either several test substances may be given (cocktail approach) or several metabolites of a single test substance may be analysed (metabolic fingerprint approach). The enzyme-specific analysis of liver function results in a preference for analysis of the metabolites rather than analysis of the clearance of the parent test substance. There are specific methods to quantify the activity of cytochrome P450 enzymes such as CYP1A2, CYP2C9, CYP2C19MEPH, CYP2D6, CYP2E1, and CYP3A, and phase II enzymes, such as glutathione S-transferases, glucuronyl-transferases or N-acetyltransferases, in vivo. Interactions based on competitive or noncompetitive inhibition should be analysed specifically for the cytochrome P450 enzyme involved. At least 5 different types of cytochrome P450 enzyme induction may result in major variability of hepatic function; this may be quantified by biochemical parameters, clearance methods, or highly enzyme-specific methods such as Western blot analysis or molecular biological techniques such as mRNA quantification in blood and tissues. Therapeutic drug monitoring is already implicitly used for quantification of the enzyme activities relevant for a specific drug. Selective impairment of hepatic enzymes due to gene mutations may have an effect on the pharmacokinetics of certain drugs similar to that caused by cirrhosis. Assessment of this heritable source of variability in liver function is possible by in vivo or ex vivo enzymological methods. For genetically polymorphic enzymes and carrier proteins involved in drug disposition, molecular genetic methods using a patient's blood sample may be used for classification of the individual into: (i) the impaired or poor metaboliser (homozygous deficient); (ii) the extensive (homozygous active) metaboliser group; and (iii) the moderately extensive metaboliser (heterozygous) group. For hepatic blood flow determinations, galactose or sorbitol given at relatively low doses may be much better indicators than the indocyanine green.(ABSTRACT TRUNCATED AT 400 WORDS)