(720c) Statistical Evaluation of Modeling Approaches of Drug Release Profiles for HPMC Matrix Tablets

Statistical Evaluation of Modeling
Approaches of Drug Release Profiles for HPMC Matrix Tablets

Wenzhao Yang¹, Jin Zhao², Jamie Curtis-Fisk², Karen Balwinski², True Rogers², and Shrikant Khot²

(1)   The Dow Chemical Company, Herbert D. Doan R&D Center, Midland, Michigan, 48674, USA

(2)   Pharmaceutical Excipients
R&D, Dow Food, Pharma & Medical, Midland, MI 48674

Matrix
tablets containing hypromellose or hydroxypropyl methylcellulose (HPMC) are
widely used in oral drug delivery modulate drug release over time.  Drug
release kinetics is primarily influenced by the kinetics and extent of polymer
swelling and erosion, and drug (or active pharmaceutical ingredient (API)) dissolution,
diffusion, and/or erosion through the polymer matrix.   Mathematical modeling
of drug release profiles has been developed in the past to enable
pharmaceutical scientists to design and optimize matrix tablet formulations ¹²³⁴. This research
focuses on using statistical analyses to compare and assess the suitability of
various models for METHOCEL™ Premium HPMC matrices.

Linear
mixed model (LMM) without empirical models and two-stage model (TSM) with
empirical models were used to develop predictive models.  Multiple regression
methods (Stepwise Regression, Partial Least Square and Generalized models) in
JMP were applied to incorporate drug property factors for modeling drug release
rate in TSM.  The dataset analyzed included API with solubility ranging from 0.2
to 1000 mg/ml and API concentration from 12.5 to 50 wt%,  HPMC grades of
METHOCEL™ K100M, METHOCEL™ K4M, METHOCEL™ K100LV and HPMC concentration from 20
to 40 wt%.  The two models were compared based on the split validation approach
using the regression coefficient (r²) and mean square error (MSE). 
The LMM has better overall prediction accuracy compared to the TSM. However,
TSM outperformed LMM in predicting drug release rate during the first two hours
following introduction to dissolution media. It has also been found that the
fitting parameters of empirical models such as Korsmeyer-Peppas and
Peppas-Sahlin models have strong correlations as shown in Figure 1. In addition
to the drug characteristics, the effect of polymer viscosity grade and API concentration
were found as key impacting factors in the model. 

The
predictive model developed here provided mechanistic insight into kinetic release
profile of HPMC matrices. This can be used to aid pharmaceutical scientists
with efficient and streamlined formulation design and in vitro drug
release experimentation to deliver desired modified-release performance.  A
more comprehensive model will be further developed to predict the effect of key
controllable variables such as type and concentrations of fillers, polymer
concentrations, and compression force.

Figure
1. Correlation of fitting parameters of Korsmeyer-Peppas Model.

References

1              J Sipemann and N  Peppas, "Mechanism of drug release from delivery
systems based on HPMC," Adv Drug Deliv Rev 48, 139 (2011).

2              N.
A. Peppas and B. Narasimhan, J. Cont. Rel. 190, 75-81 (2014).

3              LS 
Koester, GG  Ortega, P Mayorga, and VL Bassani, "Mathematical evaluation
of in vitro release profiles of hydroxyl propyl methylcellulose matrix tablets
containing carbamazepine associated to cyclodextrin.," Eur J Pharm
Biopharm 58, 177–179 (2004).

4              MC
Gohel and MK Panchal, "Novel mathematical method for quantitative
expression of deviation from the Higuchi model.," AAPS Pharm Sci Tech 1
(1-6) (2000).