Aqueous solubility and dissolution rate does not adequately predict in vivo performance: a probe utilizing some N-acyloxymethyl phenytoin prodrugs.

Some physicochemical properties of N-acyloxyalkyl prodrugs of phenytoin were reported previously.(1,2) It was shown that despite their lower aqueous solubilities relative to phenytoin, these lower-melting prodrugs with apparently disrupted crystalline structures gave either comparable or enhanced in vitro solubility and dissolution rate in simulated intestinal media made up of bile salts and lecithin (SIBLM).(2) The current objective was to compare the in vivo behavior of two of these prodrugs to phenytoin in dogs and attempt to correlate the in vitro behavior to their in vivo behavior. The oral bioavailability of phenytoin after administration of phenytoin (1) and the selected prodrugs, 3-pentanoyloxymethyl 5, 5-diphenylhydantoin (2) and 3-octanoyloxymethyl 5, 5-diphenylhydantoin (3), in fed and fasted beagle dogs were compared to intravenously administered phenytoin. Phenytoin and its prodrugs showed improvement in fed-state phenytoin bioavailability relative to the fasted state indicating that food enhanced the delivery of phenytoin from phenytoin and its prodrugs. The increased bioavailability in the fed state may be due to stimulation of bile release by food and, for the prodrugs, possible catalysis of their dissolution by lipases.(3) In both, fasted and fed states, prodrugs 2 and 3 gave higher AUC values of phenytoin than the parent compound. The enhanced bioavailability of phenytoin after oral administration were more obvious in fed dogs. Although enhanced, AUC values of phenytoin from the prodrugs relative to phenytoin were not statistically different (at 95% confidence level) in fasted state, but were different in fed state. Although the aqueous solubilities and dissolution of both prodrugs were lower than phenytoin, dissolution of 2 and 3 was equivalent and greater, respectively, relative to phenytoin in SIBLM. As expected, the in vivo behavior correlated better with the in vitro SIBLM dissolution behavior. These results indicate that aqueous solubility per se does not adequately predict in vivo behavior.