Never­theless, poisoned patients are often unnecessar­ily subjected to these potentially harmful meas­ures, and the literature is full of anecdotal accounts of miraculous recovery attributed to such treatment (Winchester et al. 1977). Prop­erly controlled clinical trials are difficult to carry out, and very few have been published. With the possible exception of forced alkaline di­uresis for poisoning with salicylate and long act­ing barbiturates such as phenobarbitone, none of these methods for enhancement of drug re­moval has ever been shown to reduce morbidity or mortality in poisoned patients (Todd 1984).

Indeed, some studies suggest the opposite result. This is not to say that such measures are never necessary, or indeed sometimes life saving, but a more critical appraisal of their role is required.

In some cases, the drug presumed to have been taken has never been chemically identi­fied, while, in others, haemodialysis has been carried out in patients with less than therapeutic plasma concentrations of the drug in question. Other studies have shown removal of only a very small and insignificant fraction of the ingested dose, sometimes amounting to the equivalent of less than 1 tablet or capsule (see, for example, Comstock et al. 1983; Heath et al. 1983). A mis­leading impression of efficacy may be gained by the use of nonspecific analytical methods for drug assay (Prescott 1974).

Other Binding Agents

Other agents have been used in attempts to bind unabsorbed drug in the gastrointestinal tract. Paraquat, a lethal weedkiller for which there is no known antidote, binds very strongly to Fuller’s earth and bentonite, and these ad­sorbents are used routinely in the treatment of paraquat poisoning.

Although bentonite may re­duce the normally slow absorption of paraquat in pure aqueous solution in rats (Smith et al. 1974), ail the evidence points to extremely rapid absorption of paraquat from commercial weed­killers in man. Our experience of paraquat poi poi­soning has been disastrous, with no apparent benefit from the early use of bentonite.

Cholestyramine binds acidic drugs and can reduce the absorption of paracetamol (aceta­minophen) taken at the same time. Like acti­vated charcoal, however, it is virtually useless when the delay between ingestion and admin­istration exceeds 1 hour (Dordoni et al. 1973).

This is because, unlike say sodium or glucose, the in­take of a drug is quite well known and the pro­cesses of distribution and elimination are usu­ally very simple and not under the control of a multitude of homeostatic controlling reflexes.

Given an accurate dosing history and one or more drug concentrations, it is possible to de­scribe the pharmacokinetic processes of distri­bution and elimination quite precisely in an individual patient and to make accurate predic­tions of concentrations at future points in time, whatever dosing regimen is used.

The ability to interpret drug concentrations and extract infor­mation so that future concentrations can be pre­dicted and a rational dosing scheme instituted, requires a different kind of intellectual effort from that needed to interpret, say, a serum glu­cose concentration. The latter is mostly inter­preted by reference to a so-called ‘normal range’ – usually the 95% confidence interval based upon measurements from a sample of a ‘normal’ population.

Under most circumstances, if the serum glucose concentration is within the “nor­mal range”, little further attention is paid to it If it is outside of the ‘normal range’, diagnostic efforts are made to determine what pathophys­iological process is disturbed. But the precise value will be used only in a semi quantitative fashion, eg. high, very high, or extremely high, or in reference to some previously defined diag­nostic threshold value; for example, diabetes mellitus may be diagnosed if the glucose con­centration is greater than 10 mmol/L.

On the other hand, all drug concentrations exceed the “normal” range because it would be abnormal to be able to detect any therapeutic substance in the serum of a normal, healthy individual.

The quantitative information provided by a drug concentration measurement can be use­fully applied with up to 2 significant figures in the determination of, say, drug clearance. This degree of precision is implicit in use of this in­formation because individual dosage decisions will often be made with 2 significant figures in the dose size.

The quantitative approach to therapeutic decision making is relatively new to the art of medicine. Most clinicians currently practicing medicine will not have been taught very much pharmacokinetics during their undergraduate training and may therefore have only a very hazy idea of the quantitative decisions that form the basis of rational therapeutics.

Furthermore, younger clinicians may be misled into thinking that the pharmacokinetics they were taught at medical school are irrelevant to modern medi­cine because their senior colleagues pay no at­tention to pharmacokinetic detail and make therapeutic decisions in a seemingly capricious fashion. Such a conclusion may be quite false because therapeutic decisions made by an ex­perienced clinician are founded upon a wide base of knowledge gained from treating many similar patients.

This prior knowledge of the character­istics of the population being treated provides an empirical, but nevertheless frequently satis­factory, guide to making an appropriate quan­titative and qualitative therapeutic decision. Recognition of the value of such prior infor­mation when faced with an individual patient about whom little is known is the basis of the latest techniques of rational, quantitative phar­macokinetic and pharmacodynamic forecasting.

For the person who must make a therapeutic choice on behalf of an individual patient, the application of quantitative pharmacokinetic principles can substitute for the advice of a more experienced colleague. For all clinicians, young and old, these same principles can be applied to new therapeutic entities and reduce the suf­fering of patients who would otherwise be ex­posed to the vagaries of a trial and error ap­proach.