Objectives. The amounts of fluoride released by different glass ionome
r formulations were compared on the basis of individual fluoride relea
se profiles in order to derive the effect of the physical and chemical
formulation on the fluoride release process. Methods. The fluoride re
lease profiles of each of five specimens of ten glass ionomer cements
setting by an acid-base reaction were investigated. The profiles were
obtained by determining the amount of fluoride released [F] after equi
librating the samples at 37 degrees C in distilled water for 140 d. Th
e {[f],t}-profiles were compared with a Multivariate Data Analysis on
the basis of a Principal Component Analysis. Results. The Multivariate
Data Analysis reveals that eight of the ten glass ionomers can be cla
ssified into four distinct groups. When the cumulative amount of fluor
ide released by each sample, [F](c), is calculated and fitted as a fun
ction of time, a regression analysis (r > 0.99) reveals that [F](c) fo
r all samples is most adequately represented by [F](c) = ([f](l).t)/(t
+t(1/2)) + B .root t, indicating that two kinetic processes are respon
sible for the fluoride release profiles. Significance. A comparison of
the parameters of this equation shows that the physicochemical ration
ale for the classification of the glass ionomers conforms to differenc
es in the kinetics of these processes which are determined by the qual
itative and quantitative chemical composition as well as by the presen
tation (hand mixed vs. capsules) of the glass ionomer. From the classi
fication, it becomes apparent that different formulations can result i
n the same fluoride release profiles.