Development and evaluation of a population pharmacokinetic model for cilobradine, an If channel blocker
G. Fliss (1), A. Staab (2), C. Tillmann (2), D. Trommeshauser (2), H.G. Schaefer (2), C. Kloft (1)
(1) Dept. Clinical Pharmacy, Institute of Pharmacy, Freie Universitaet Berlin, Berlin, Germany (2) Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach a.d.R. Germany
Objectives: The If channel blocker cilobradine belongs to a class of bradycardic agents selectively decreasing heart rate by reducing the diastolic depolarisation rate in the sinus node. Hence, cilobradine might be beneficial in the treatment of cardiovascular diseases, e.g. ischemia. The objective of this study was to evaluate the population pharmacokinetic (PopPK) characteristics based on data of 6 clinical phase I trials with different formulations and to assess the predictive performance of the model developed.
Methods: Single doses of 1.25-40 mg cilobradine were administered as p.o. solution, p.o. capsule or 20 min i.v. infusion. The capsule was also given once daily (0.6-20 mg) for 7 or 15 days. PK profiles of 162 males with 2733 plasma samples were analysed (development data set). NONMEM, version V, level 1.1, with the FOCE interaction method was used for data fitting and assessment of several covariates by forward inclusion and backward deletion techniques. Model evaluation was performed using a new data set (evaluation data set) including 1713 plasma concentrations of p.o. solution over a dose range of 0.25-5 mg.
Results: The data were best described by a 3-compartment
model with first-order absorption and elimination. The first distribution
process revealed administration route-dependent characteristics; after i.v.
dosing the initial distribution phase was faster than after p.o dosing.
Therefore, V2, V3 and Q3 were separately estimated for i.v. or p.o. data.
Typical Vsspo and Vssiv were large (95.8 L and 130.3 L, resp.), CL was 21.5 L/h
and Q3 15-fold higher for i.v. than for p.o. data (99.8 L/h vs. 6.61 L/h).
Absolute bioavailability was significantly lower for p.o. solution than for p.o.
capsule (34% vs. 43%); the capsule showed an additional lag time of 0.154 h.
Covariate analysis revealed a statistically significant relation between KA and
dose. It was best described by a positive saturation function resulting in KAmax
of 0.43 h-1 which was nearly reached at the dose of 5 mg, the dose at 0.5 KAmax
was 1.00 mg, i.e. the relation was primarily acting in the low dose range.
Interindividual variability estimated for KA, CL, F1 and V2iv was moderate (15%
to 46%), residual variability was 26% (proportional random effect model).
Imprecision of estimates was generally low with relative standard errors (RSE)
<25% except for lag time (RSE: 52%).
The estimates for the evaluation data
set based on the final PopPK model were very similar to those of the development
data set except of the covariate relation which was not supported by the
evaluation data set. Simulations (n=500) of the evaluation data set based on the
final PopPK model but without the covariate relation revealed that almost all
observed concentrations of the evaluation data set were covered by the 90%
prediction interval of the simulated concentrations.
Conclusions: A PopPK model has been successfully developed describing the plasma concentration-time course of cilobradine after administration of different formulations. As the covariate relationship found in the development data set was based on a limited number of data in the low dose range and could not be confirmed in the evaluation dataset it should be revisited in a larger population, preferentially in the target patient population. The PopPK model was suitable to sufficiently predict concentrations of a different study design. Therefore, the model can serve as a tool to simulate and evaluate different dosing regimens for further clinical trials.