An assessment of prediction accuracy of two IVIVC modelling methodologies.
Clare Gaynor (1), Adrian Dunne (1) and John Davis (2)
(1) UCD School of Mathematical Sciences, University College Dublin, Belfield, Dublin 4, Ireland (2) Clinical Pharmacology, Pfizer Global Research and Development, Sandwich, England
Introduction: In Vitro - In Vivo Correlation (IVIVC) models are extensively used in the drug development process. The accuracy with which In Vitro observations can be used to predict an In Vivo response is paramount and may depend on the choice of method used to develop such a model. Two approaches to developing these models are predominantly used: a traditional deconvolution-based method and an alternative convolution-based technique. In spite of the fact that there are are a number of areas of concern regarding the deconvolution-based methods [1], they are routinely used. The convolution-based alternative [2],which does not suffer from the same weaknesses is, however, relatively rarely employed. It has previously been shown [3] that, where an IVIVC relationship does exist, a deconvolution-based method is more likely than a convolution-based procedure to fail the FDA validation test [4] (32% and 0.1% failure rates respectively). The aim of this study is to supplement these results by further investigating the performance of both a deconvolution and a convolution based method and quantifying any difference in accuracy of prediction.
Methods: In Vitro and In Vivo (extended release and reference) data were simulated for four formulations of a drug according to a model which incorporated an IVIVC relationship. Each subject's reference data was analysed separately using the ADVAN2 subroutine provided in the NONMEM software developed by Beal and Sheiner [3]. The dataset was then analysed twice to produce two predicted plasma concentration-time profiles. First the CoDe deconvolution method proposed by Hovorka et al [6, 7] was implemented. This was followed by the convolution based technique - a modified version of that reported by O'Hara et al [2] which implemented a custom written PRED subroutine for NONMEM. Each method's predictions were used to calculate the peak plasma concentration value and area under the plasma concentration curve for each formulation and were compared to the values calculated from the simulated data to calculate percentage prediction errors.
Results and Conclusions: The convolution method produces an IVIVC model that comfortably meets the FDA criteria. The deconvolution based method not only fails to produce a satisfactory model but the shape of the predicted plasma concentration curves do not match the data. Together with previously reported results [3], these findings provide further support for the use of the convolution based approach.
References:
[1] Dunne A., Gaynor C. and Davis J. (2005) Deconvolution Based Approach for Level A In Vivo-In Vitro Correlation Modelling: Statistical Considerations. Clinical Research and Regulatory Affairs, 22, 1-14.
[2] O'Hara T., Hayes S., Davis J., Devane J., Smart T. and Dunne A. (2001) In Vivo-In Vitro Correlation (IVIVC) Modeling Incorporating a Convolution Step. Journal of Pharmacokinetics and Pharmacodynamics, 28, 277-298.
[3] Gaynor, C. Dunne, A.and Davis, J. (2006) Comparison of conventional In Vitro - In Vivo Correlation methodology with non-linear mixed effects modelling. PAGE 15 2006 Abstr 953 [www.page-meeting.org/?abstract=953]
[4] Food and Drug Administration (1997) Guidance for Industry: Extended Release Oral Dosage Forms: Development, Evaluation, and Application of In Vitro/In Vivo Correlations.
[5] S.L. Beal and L. B. Sheiner. NONMEM User's Guides, NONMEM Project Group, University of California, San Francisco, 1992.
[6] Hovorka, R., Chappell, M.J., Godfrey, K.R., Madden, F.N., Rouse, M.K. and Soons, P.A. (1998) CODE: A Deconvolution Program Implementing a Regularization Method of Deconvolution Constrained to Non-Negative Values. Description and Pilot Evaluation. Biopharmaceutics and Drug Disposition 19, 39-53.
[7] Madden, F. N., Godfrey, K.R., Chappell, M.J., Hovorka, R. and Bates, R. A. (1996) A Comparison of Six Deconvolution Techniques. Journal of Pharmacokinetic and Biopharmaceutics, 24: 3, 283-299.
