A new approach to the study of serum concentrations of orally administered cephalexin

Abstract

Studies of cephalexin in man showed that the conventionally constructed average serum level curves could give misleading results, with the average peak serum level from the curve being lower than the peak titres of the individual curves contributing to it. It was also difficult to determine whether there were significant differences between such curves obtained from different preparations. Smooth curves were constructed from the observed data by computer methods; they were then arranged by the computer to have their peaks coincident in time and then averaged to give a curve much more similar in shape to its components. After the computer had fitted the smooth curve for each individual, it was possible to calculate a number of different parameters for each volunteer. Analysis of variance could then be done on these parameters, and thus significant differences between the results from tests on different preparations could readily be determined.