Carl M. Metzler

Efficacy and residual effect evaluation of a new hypnotic, triazolam

In all animal studies, triazolam had a higher relative potency than both diazepam and chiordiazepoxide. Triazolam was less toxic than diazepam, indicating it may well have an exceedingly large therapeutic ratio. Initial studies in man indicated that the main pharmacological effects of triazolam were sedation and sleep induction, and this led to the decision to study this compound as a hypnotic agent.1 One problem with presently available hypnotic agents is the hangover or resid-

Usefulness of the Two-Compartment Open Model in Pharmacokinetics

Abstract Compartment models are widely used in pharmaceutical research to quontitate the kinetics of absorption, distribution, metabolism and excretion of a drug. In this article two applications of the two compartment open model are discussed. This discussion shows that, in spite of the extreme simplification and obvious weaknesses of this model, it has properties that make it a useful tool in drug research when properly used.

Effect of H2 blockade in the challenged allergic nose

Abstract To examine whether H 2 blockade might prevent development of allergen-induced nasal congestion, we subjected 18 volunteers with allergic rhinitis to nasal allergen challenge after pretreatment with cimetidine or with placebo. Response after cimetidine was more severe than that which followed placebo. H 2 blockade has a subtle adverse effect on nasal allergic response.

Linear and nonlinear system approaches in pharmacokinetics: How much do they have to offer? II. The response mapping operator (RMO) approach

The description of the relationship between different responses measured simultaneously in the same subject is commonly described in terms of specific pharmacokinetic models such as linear compartmental models. An alternative system approach involving response mapping operators (RMOs) is presented. The theoretical and mathematical basis of the RMO approach are derived. The assumptions, limitations, and practical significance of the RMO approach are discussed. The derivation of the RMO is illustrated with several examples. An algorithm and computer program for implementing the RMO in a routine manner is presented. The usage of the computer programs RMO and MAP presented are demonstrated using the pharmacokinetics of trimazosin and cefamandole in humans as examples. The RMO approach offers a new and powerful tool in pharmacokinetic analysis, which allows the investigator to approach certain problems in an objective, rational way without getting involved in specific pharmacokinetic modeling.

Effect of flurbiprofen, a cyclooxygenase inhibiting drug, on induced allergic rhinitis

To test whether prostaglandins contribute to the nasal allergic reaction, we subjected allergic rhinitis patients to nasal allergen challenge after treatment with flurbiprofen, a cyclooxygenase inhibitor. We challenged 18 patients after two pretreatment doses of flurbiprofen, 75 mg, chlorpheniramine, 6 mg, both drugs combined, and placebo. All patients received all treatments. None of the treatments affected measured nasal airway resistance. Nasal secretion weight, number of sneezes, and overall subjective severity scores for all active treatments differed significantly from placebo treatment. Flurbiprofen proved nearly as effective as chlorpheniramine in reducing severity of induced rhinitis. Some mechanism that can be influenced by flurbiprofen (possibly prostaglandin synthesis) contributes to the acute, induced allergic reaction.

The efficient use of NONLIN for unbalanced multiple dose data.

One of the uses of pharmacokinetic modeling is the fitting and prediction of blood levels after multiple dosing. The equations that govern this process are readily available for the case of equally spaced dosing intervals. For unequally spaced intervals, some rather inefficient methods have been suggested for use with NONLIN (1). This article shows how to implement parameter estimation and fitting of such data by means of NONLIN by summing the appropriate single-dose equations.

Statistical Properties of Kinetic Estimates

Most kinetic parameters of interest are related to the observed data by nonlinear mathematical expressions. Thus the statistical properties of the estimates are influenced by many factors: The model and its degree of nonlinearity, the number of observations, the values of the independent variable, the error structure, the method of estimation, transformations of the parameter space, and the initial estimates if the estimation procedure is iterative. For most models, an asymptotic estimate of the covariance matrix can be written which includes some of these factors. The validity of the asymptotic estimate is questionable in the usual pharmacokinetic study with few observations. Some small-sample properties of the estimates can be inferred from Monte Carlo simulations, but the vast variety of possible situations limits the usefulness of this approach. A review of some Monte Carlo studies indicates that for many pharmacokinetic situations the bias of the estimates is small, but the estimates of variability are poor. In general, estimates of the parameters of linear systems have small-sample properties that are well approximated by the asymptotic theory. Parameter estimates from nonlinear systems are more difficult to compute and the statistical properties for small samples are not well approximated by asymptotic theory.

Clindamycin bioavailability from clindamycin-2-palmitate and clindamycin-2-hexadecylcarbonate in man

To determine the bioavailability of clindamycin from the microbiologically inactive clindamycin-2-palmitate and clindamycin-2-hexadecylcarbonate in man, a three-way crossover study was conducted with oral administration of the two esters and clindamycin hydrochloride. In each case, serum clindamycin bioactivity concentrations were fitted to a one-compartment open model with an initial lag time. Analysis of variance of measured quantities (serum concentrations and urinary excretion) and of derived pharmacokinetic parameters showed that for every comparison except maximum serum concentrations clindamycin-2-palmitate was not significantly different (at p=0.05) from clindamycin hydrochloride. Clindamycin-2-hexadecylcarbonate gave significantly different values from those for the hydrochloride in all cases except the rate constant and half-life for elimination from the serum. The palmitate was the superior ester and was bioequivalent to the hydrochloride in man.

Bioavailability/Bioequivalence: Study Design and Statistical Issues

In this discussion of bioavailability and bioequivalence studies only those studies in which the observations concern concentrations of drug in blood (or one of its components) or amounts of drug in urine collections will be discussed. Much more use could be made of pharmacologic responses in evaluating and comparing bioavailabilities, but that is a different topic. The discussion is also limited to bioavailability of drug from conventional, rapid-release formulations. Most of the discussion may also apply to controlledor sustained-release products, but until there is a better understanding and agreement on the relevant aspects for comparing these formulations, it is not clear what additional statistical considerations may be needed. Many pharmacologic-pharmaceutical concerns may impact the design and analysis of these studies, for example: the relationship of blood levels to pharmacologic effect; normal subjects or patient; singleor multiple-dose studies for a drug that will be used chronically; bioavailability of metabolites. These are not statistical questions (although statistics may help answer them) and are not discussed here. It is the nature of biologic and clinical research that it is not possible to truly replicate the system being studied, although stable isotopes offer some potential here. Because there is so much variability among the systems or individuals being studied, it is usually advantageous to study each individual under more than one of the conditions of the trial. This has the potential for introducing correlation in the error structure. How this possible correlation affects design and analysis will be discussed later. Presented here are general concepts without equations or formulas and with references for technical discussions and details of analysis. It is under-

Practical ENT. Incidence of palpable cervical nodes in adults.

Palpable cervical nodes were found in 56% of 561 adult outpatients studied. The incidence decreased with advancing age. Only two patients with palpable nodes had a history of neck cancer.