Limitations of the maximum entropy principle in devising drug input rate.

A computer program applying the principle of maximum entropy to the analysis of drug absorption rate has been developed. Plasma concentrations of amoxicillin obtained after oral and intravenous dosing have been analysed, together with simulated data corresponding to a complex input. Amoxicillin absorption rates devised by the program were similar to those obtained by a standard deconvolution method, although they were displayed as an almost continuous profile. However, improbable fluctuations were obtained with some data sets and the fraction absorbed was underestimated by 13%. With the simulated data, the maximum entropy program did not provide a better solution than the standard deconvolution procedure, and it was sensitive to the addition of random error and to the number of samples. The maximum entropy principle, as implemented in our computer program, may not have a better performance than standard deconvolution procedures, especially in human experiments where the number of blood samples is usually limited.

Diuretic effect and diuretic efficiency after intravenous dosage of frusemide.

1. Frusemide was given intravenously at a dose of 5 mg kg-1 to five healthy volunteers and the diuresis was assessed by frequent spontaneous voiding over 5 h. Urinary volume and contents of sodium, chloride, potassium and frusemide were measured. 2. Diuretic response was evaluated using the sigmoid Emax model and non linear regression of diuresis vs frusemide excretion rate. The time courses of diuresis (pharmacological effect) and diuretic efficiency were constructed from the fitted parameters of the sigmoid Emax model. 3. The frusemide excretion rate associated with maximum efficiency was found, as predicted theoretically, to be less than the excretion rate associated with 50% of maximum effect in four of the five subjects in whom the slope factor was less than 2. 4. The effect over time is dependent both on the instantaneous drug effect but also on its pharmacokinetic properties and mode of administration. An intravenous bolus is the least efficient mode of administration while a controlled input producing a frusemide excretion at maximum efficiency should yield up to a 2.3 times higher diuretic response.