Drug Kinetics

1 Introduction

The review period 1978 — 1979 has witnessed a continued increase in the literature of many aspects of drug kinetics and, as in previous volumes, this chapter presents the literature concerning some of the major drug classifications. Stringent space restrictions require a selective rather than a comprehensive treatment, with emphasis being placed on fewer drug groups and those communications which, in the reviewer's opinion, reflect the more important contributions in particular areas.

The increasing awareness of the importance of factors influencing drug absorption and disposition is reflected in the introduction of two new journals, Biopharmaceutics and Drug Disposition and Pharmacy International, while a new journal concerned with nutrient-drug intcractions will appear shortly. New books have been published on the pharmacokinetics of chemotherapeutic agents, principles and perspectives in drug bioavailability, drug fate and metabolism, and the fate of drugs in the elderly. A useful text on basic pharmacokinetic principles has been introduced by Niazi.

Theoretical papers published during the review period include discussions on the time-dependency of the induction of drug metabolism, the use of statistical moments in pharmacokineticanalysis, and the fitting of kinetic data to biexponential functions. This last publication compares the use of expressions of a form of equation 1

C = Ae-αt + Be-βt (1)

and also the differential and integrated forms of equation 2

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

for the non-linear computer analysis of biexponential data. In equation 1, A, B, α, and β are model independent hybrid constants, while in equation 2 the k's are first-order rate constants, and Xp and Ic are model-related parameters. The authors show that while computer convergence is slower with the differential form of equation 2 than with the integrated form, the accuracy and precision obtained from equations 1 and 2 are similar. They point out, however, that the use of either the differential or the integrated form of equation 2 is necessary to obtain the variance of the least-squares estimate of the particular k values.

Problems associated with drug accumulation in the body during continuous or repetitive doses continue to attract attention, and methods are described to predict accumulation profiles for drugs obeying Michaelis–Menten kinetics following zero-order, first-order, and instantaneous absorption, and also for drugs with nonlinear protein and tissue binding characteristics. Estimates of kinetic parameters for drugs exhibiting linear kinetics, based on two data points obtained at steady-state, are described by Bjornsson and Shand. The method, which is based on simple algebraic functions, is more reliable for drugs obeying one-compartment kinetics than for those with more complex kinetic profiles. A previously described method for rapid attainment of steady-state drug levels using a two-step infusion has been expanded by Zimmerman, who also describes the contribution of error effects in the prediction of steady-state values for drugs with biexponential drug profiles. Similar error effects have been described in predicting drug and metabolite urinary excretion profiles by the accelerated convergence technique for drugs which obey biexponential kinetics.

Applications of numerical deconvolution and linear systems analysis have been described for the calculation of in vivo drug dissolution, absorption, and metabolism, and also pharmacological data. Although theoretically sound, these approaches may suffer from the common problem of function identity and also data noise.

Other communications, of perhaps more practical application, have considered the dependence of pharmacokinetic parameters on body weight, and theoretical aspects of drug monitoring in serum and saliva. The use of general pharmacokinetic principles in drug therapy is advocated by Boëthius and Sjöqvist. These authors comment on the considerable variation in drug dosage, particularly of antidepressants and β-receptor blocking agents, and the poor correlations in some cases between dose adjustment and patient condition. A series of articles on the use of pharmacokinetics in drug therapy has been published by Schumacher et al. A method of individualization of drug dosage, with particular reference to chloramphenicol and theophylline, has been described by Koup et al. This method, which is based on a single drug-level determination, is as accurate as traditional pharmacokinetic methods requiring more extensive drug-level data. Supporting evidence for an earlier contention, that the use of one-compartment model kinetics may introduce error into the calculation of drug clearance values for drugs which obey more complex models, is presented by Dvorchik and Vesell. From previously published data, these authors show that the calculated error varied from 1% in the case of phenobarbital, to 81% in the case of cephalexin, and 196% in the case of ampicillin.

Two authors have commented on the errors involved in the use of the trapezoidal rule for determining areas under drug plasma concentration time curves. It is recommended that incorrect area estimates may be avoided if a linear trapezoidal method of the form of equation 3 is employed for pre-peak and plateau data in absorption studies, while a logarithmic trapezoidal method of the form of equation 4 is used for post-peak data:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

The latter equation, in which Ci is concentration at time ti, applies for all exponentially declining drug profiles.

Recent advances in the use of physiologically based pharmacokinetic models, with particular reference to thiopental cytarabine, actinomycin D, sulphobromophthalein, salicylate, lidocaine, adriamycin, digoxin, methotrexate, and cephalosporins are reviewed by Himmelstein and Lutz....