Mean input times of three oral chlorprothixene formulations assessed by an enhanced least-squares deconvolution method

The efficacy and quality of drug formulations are determined mainly by their bioavailability, which is defined by the rate and extent of drug absorption. The zero and first moments of the serum concentration time profile provide relevant information on the bioavailability. On the basis of the body residence time distributions, mean input times may be used as drug absorption rate parameters, but due to computational errors the statistical moments procedure is in some instances of limited value. To circumvent these problems we have developed a procedure to calculate mean input times from input profiles obtained by least-squares deconvolution. We enhanced the performance of the deconvolution method by directly generating initial estimates of one input rate for each sampling interval and compared the statistical properties of various input rate characteristics using data from a bioavailability study on four chlorprothixene preparations. The analysis of variance revealed that estimates of mean input times depended on the calculation procedure. Mean input times estimated by the least-squares deconvolution method were more reliable and less variable than those computed as differences of mean body residence times.