Aarons, L. and Gupta, S. K
Pharmacy Department, University of Manchester, UK.
A longitudinal clinical trial was undertaken in which verapamil was administered orally daily at 10 pm to patients for a period of 12 weeks. The doses of verapamil used were placebo, 120, 180, 360 and 540mg, in a number of different sequences, 103 subjects completed the study. Peak and trough blood samples were taken on week 12 and continuous 24hr continuous ambulatory blood pressure monitoring was performed. The plasma concentration data was modelled with a one compartment model in which the input was determined by the delivery characteristics of the device coupled to a first order absorption mechanism. The kinetics were dose dependent with apparent CL for the r-enantiomer varying from 96.6 (+/-9.7) L/hr at the 120mg dose to 58.4 (+/-5.3) L/hr at the 540mg dose. Apparent volume of distribution showed no clear trend with dose, varying between 494L to 712L. there was considerable interindividual variability in CL (43.3%) and V (17.6%).
Systolic blood pressure was related to r-verapamil plasma concentration, after baseline correction, using an Emax model. Several methods of baseline correction were tried. Simple point-by-point subtraction of baseline values from the corresponding times gave an adequate description of drug effect. The use of a population longitudinal spline with only shift and scale parameters proved not to be sufficiently flexible to describe the baseline profile and it was impossible to fit the Emax model to this data. Individual cubic spline interpolation gave the best representation of the data. The C50 estimates for the Emax model was 455 ng/L with an interindividual variability of 32%. The regression of AUC of effect against AUC of concentration, also using an Emax model, gave a much more stable analysis. The corresponding C50 estimate from this analysis, 1220 ng/L, obtained by taking a 24hr average of the AUC50 was quite different from the previous estimate based on the concentration data. In conclusion, several methods for baseline correction were used and results were compared. Point-by-point method proved to be the easiest to implement and provided comparable results.
Reference: PAGE 6 () Abstr 652 [www.page-meeting.org/?abstract=652]
Poster: poster