Experimental design application and interpretation in pharmaceutical technology - Computer-Aided Applications in Pharmaceutical Technology - page 49

liposomal formulations (Loukas, 2001). These formulations
contain ribofl avin in either free form or complexed with cyclodextrin
as a model drug, sensitive to photochemical degradation, as well
as various light absorbers and antioxidants incorporated into the
lipid bilayer and/or the aqueous phase of liposomes. The following
factors were varied on 2 levels: ribofl avin complexation (free
vitamin or γ -cyclodextrin complex), presence or absence of light
absorbers oil red O, oxybenzone, dioxybenzone, sulisobenzone, and
the antioxidant β -carotene. The multilamelar liposomes were
prepared either by the dehydration-rehydration technique or by the
disruption of lipid fi lm method containing cholesterol in low or high
concentrations, 1,2-distearoyl- sn -glycero-3-phosphocholine (DSPC)
as an alternative lipid, and sonicated through a bath or probe
sonication for a low or higher period of time. All these variations
comprise the 11 factors that directly affect the physical stability of
liposomes, as well as the chemical stability of the entrapped
vitamin. To perform a full factorial design for the examination of
11 factors, at 2 levels for each, it would be necessary to prepare
2 11 = 2048 liposomal formulations. The Plackett-Burman design
allows reduction in the number of experiments from 2048 to 12,
for 11 factors studied, where each is at its 2 levels. As described
in the introduction, the Plackett-Burman design is a 2-level
fractional factorial design. The effect of each factor in the presented
study can be determined as:

￿

￿ ￿

Effect = 1/6 [ Σ ( y at + level) − Σ ( y at − level)] [3.9]

The authors have highlighted that the Plackett-Burman design is
especially useful for a multivariate system with many factors that
are potentially infl uential on system properties. Once these
infl uential factors are recognized (by Plackett-Burman design) and
their number signifi cantly reduced, other forms of experimental
design, such as full or fractional factorial design, are used.
Plackett-Burman designs cannot be used for detection of factors
interactions. It was found that the presence of light absorber oil
red O demonstrates the most signifi cant effect on liposome