Optimisation of
Ondansetron Orally Disintegrating Tablets Using
Artificial Neural Networks
Buket Aksu1*,
Gizem Yegen2, Sevim Purisa3, Erdal
Cevher2 and Yıldız Ozsoy2
1Santa Farma Drug
Pharmaceuticals, Boruçiçeği Sokak, Şişli-Istanbul,
2Department of Pharmaceutical Technology, Faculty
of Pharmacy, 3Department of Biostatistics,
Faculty of Medicine, Istanbul University, Istanbul,
Turkey
*For correspondence:
Email:
baksu@santafarma.com.tr; Tel:
+902122206400; Fax: +902122225759
Received: 24 April 2014
Revised accepted: 29 July
2014
Tropical
Journal of Pharmaceutical Research, September 2014;
13(9): 1374-1383
http://dx.doi.org/10.4314/tjpr.v13i9.1
Abstract
Purpose: To investigate the impact
of critical quality attributes (CQAs) and critical
process parameters (CPPs) on quality target product
profile (QTPP) attributes of orally disintegrating
tablet (ODT) containing ondansetron (OND) using two
artificial neural network (ANN) programs.
Methods: Different amounts of two
different commercial superdisintegrants commonly used in
ODT formulations (Ludiflash® and Parteck®) were examined
as CQAs, while three different tablet-pressing forces
were evaluated as CPPs for an orally disintegrating
tablet (ODT) formulation. The impact of CQAs, and CPPs
on the target product profile (tablet hardness,
friability and disintegration time) were analysed using
gene expression programming (GEP) and neuro-fuzzy logic
(NFL) models.
Results: NFL model defined the
relations between CQAs, CPPs and QTPP, while GEP model
favoured the use of an ODT formulation with suitable
QTPP features which contained 4 mg ondansetron, 21.30 mg
Parteck®, and 119 mg Avicel, fabricated with a
compression force of 515 psi. In this regard, the tablet
formulation demonstrated the required specifications.
Conclusion: ANN programs are a useful
tool for research and development (R&D) studies in the
pharmaceutical industry and the use of ANNs can be
beneficial in terms of raw materials, time and cost, as
demonstrated for ondansetron ODT tablets.
Keywords: Ondansetron, Critical
quality attributes, Critical process parameters, Quality
target product profile, Gene expression programming,
Neuro-fuzzy logic, Artificial neural network
