Format

Send to

Choose Destination
Drug Dev Ind Pharm. 2016;42(3):389-402. doi: 10.3109/03639045.2015.1054831. Epub 2015 Jun 11.

Combined application of mixture experimental design and artificial neural networks in the solid dispersion development.

Author information

1
a Department of Pharmaceutical Technology and Cosmetology, Faculty of Pharmacy , University of Belgrade , Belgrade , Serbia and.
2
b Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-University Duesseldorf , Duesseldorf , Germany.

Abstract

This study for the first time demonstrates combined application of mixture experimental design and artificial neural networks (ANNs) in the solid dispersions (SDs) development. Ternary carbamazepine-Soluplus®-poloxamer 188 SDs were prepared by solvent casting method to improve carbamazepine dissolution rate. The influence of the composition of prepared SDs on carbamazepine dissolution rate was evaluated using d-optimal mixture experimental design and multilayer perceptron ANNs. Physicochemical characterization proved the presence of the most stable carbamazepine polymorph III within the SD matrix. Ternary carbamazepine-Soluplus®-poloxamer 188 SDs significantly improved carbamazepine dissolution rate compared to pure drug. Models developed by ANNs and mixture experimental design well described the relationship between proportions of SD components and percentage of carbamazepine released after 10 (Q10) and 20 (Q20) min, wherein ANN model exhibit better predictability on test data set. Proportions of carbamazepine and poloxamer 188 exhibited the highest influence on carbamazepine release rate. The highest carbamazepine release rate was observed for SDs with the lowest proportions of carbamazepine and the highest proportions of poloxamer 188. ANNs and mixture experimental design can be used as powerful data modeling tools in the systematic development of SDs. Taking into account advantages and disadvantages of both techniques, their combined application should be encouraged.

KEYWORDS:

Artificial neural networks; d-optimal mixture experimental design; physicochemical characterization; poorly soluble drugs; solid dispersions; solvent casting

PMID:
26065534
DOI:
10.3109/03639045.2015.1054831
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Taylor & Francis
Loading ...
Support Center