A Time Scaling Approach to Develop an In Vitro-In Vivo Correlation (IVIVC) Model Using a Convolution-Based Technique
Cian Costello (1), Stefaan Rossenu (2), An Vermeulen (2), Adriaan Cleton (2), Adrian Dunne (1)
(1)UCD School of Mathematical Sciences, University College Dublin, Belfield, Dublin 4, Ireland. (2)Clinical Pharmacology, Advanced PK/PD Modelling and Simulation, J&J Pharmaceutical Research and Development, a Division of Janssen Pharmaceutica N.V., Turnhoutseweg 30, B-2340 Beerse, Belgium.
Objectives: In vitro-in vivo correlation (IVIVC) models prove very useful during drug formulation development, in support of dissolution specification setting and for bio waiver applications following post approval changes. Our aim in this study was to develop a population IVIVC model for drug formulations for which in vitro drug dissolution occurs over a much shorter time scale than the in vivo drug dissolution.
Methods: It is common practice to fit an IVIVC model to deconvoluted data rather than to the raw plasma drug concentration-time data because the deconvoluted data are believed to reflect the dissolution of the drug in vivo. However, the deconvolution step and the associated methodology pose some statistical concerns.1 This has motivated the development of a convolution-based population approach, which addresses the aforementioned statistical issues.2 Recent studies confirm that this approach is superior to the deconvolution method.3 Due to the disparity of the in vitro and in vivo time scales for some drugs, traditional methods of finding an IVIVC model may be unsuccessful. It is suggested that a time scaling approach be applied in order to establish an IVIVC model. We applied this methodology for a drug with these characteristics. The data were longitudinal in nature with considerable between subject variation present in the in vivo data. Using the NONMEM package, a nonlinear mixed effects model was fitted to the data with a time-scale model linking the in vitro and in vivo components.
Results: The population IVIVC model was successfully fitted. The goodness of fit of the model was assessed by comparing predicted with observed in vivo plasma concentration-time profiles. The model predicted AUC and Cmax were compared with the data based values and show differences of at most 11.6% and 8.5% for AUC and Cmax respectively. These values satisfy the FDA validation criteria for both internal and external predictability.4
Conclusions: Our study demonstrates that when attempting to fit an IVIVC model to data with a significant disparity between the in vitro and in vivo dissolution time scales, a time scaling approach can be useful. Applying a time scaling approach in conjunction with the convolution based methodology demonstrates the versatility of the methodology, and the fact that the model predictions met the FDA validation criteria is evidence that these types of model are capable of producing accurate predictions, which is always of prime importance.
References:
[1] Dunne A, Gaynor C, 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, Dunne A. (2001) In Vivo-In Vitro Correlation (IVIVC) Modelling Incorporating a Convolution Step. Journal of Pharmacokinetics and Pharmacodynamics 28:277-298.
[3] Gaynor C, Dunne A, Davis J. (2008) A Comparison of the Prediction Accuracy of Two IVIVC Modelling Techniques. Journal of Pharmaceutical Sciences in press. Published online in Wiley InterScience DOI 10.1002/jps.21220.
[4] Food and Drug Administration. (1997) Guidance for Industry: Extended Release Oral Dosage Forms: Development, Evaluation and Application of In Vitro/In Vivo Correlations.