Farkad Ezzet
Ciba-Geigy, CH-4002 Basel, Switzerland
Valsartan is one of a new class of anti-hypertensive agents, known as angiotensin II antagonists. Its pharmacokinetics and variability, both between and within subjects, are described. The data used in this investigation are plasma concentrations following oral intake of Valsartan by healthy volunteers and by hypertensive patients.
In an earlier absolute bioavailability trial in twelve healthy subjects, ValsartanÃs elimination was easily characterized as bi-phasic, however, the process of absorption could not be fully described. This is due to 1) the extra complexity of the absorption process of Valsartan, 2) limited data within the first few hours after drug intake where most of absorption takes place, and 3) high variability between subjects and in measurement error. Despite these problems, using the much larger data set, a pool of 118 concentration profiles in volunteers and fitting a two compartment model, adequate approximation of absorption was possible by either I) a zero order process or ii) a first order process with a lag time between drug intake and onset of absorption. The drawback of I) is that it has little biological support, and of ii) is that lag time is estimated to be longer than what early concentration data may suggest. The impact of the different approximations of absorption is noticed on the estimates of the elimination and transfer rates. As regard to the between subject variability, the two approximations are in close agreement.
In a phase III trial, 96 hypertensive patients provided four concentration data points at 0, 2, 4 and 6 hours following intake of a single capsule of Valsartan daily on days 1, 14 and 28 of treatment. Since this limited number of data points per dose per individual allow very restrictive models, the one compartment model with the zero order absorption was considered and fit to the steady state data of days 14 and 28. A better fit to the data was obtained by allowing the fraction of dose absorbed to vary between patients as well as to have different values at different days but constrained to average to the previously estimated population bioavailability fraction. Various goodness of fit criterion suggested adequacy of the fit, however, a further validation step was employed by super-imposing concentration data from day 1 on a 90% predictive interval obtained from the model. This has resulted in a 87% coverage, confirming the adequacy of the fit.
Although a one-compartment model with zero order absorption is an obvious simplification of the true underlying pharmacokinetics of Valsartan, this approximation would not invalidate conclusions with respect to between subject variability and subgroup differences. On subgroups, clearance appears to be reduced with age; for instance, clearance for a 45 year old patient is estimated as 1.9L/h compared to 1.3L/h for a 65 year old patient.
The above analyses were carried out using NONMEM and NLME.
Reference: PAGE 5 (1996) Abstr 576 [www.page-meeting.org/?abstract=576]
Poster: poster