Development of a Physiologically Relevant Population Pharmacokinetic in Vitro-in Vivo Correlation Approach for Designing Extended-Release Oral Dosage Formulation.

Establishing a level A in vitro-in vivo correlation (IVIVC) for a drug with complex absorption kinetics is challenging. The objective of the present study was to develop an IVIVC approach based on population pharmacokinetic (POP-PK) modeling that incorporated physiologically relevant absorption kinetics. To prepare three extended release (ER) tablets of loxoprofen, three types of hydroxypropyl methylcellulose (HPMC 100, 4000, and 15000 cps) were used as drug release modifiers, while lactose and magnesium stearate were used as the diluent and lubricant, respectively. An in vitro dissolution test in various pH conditions showed that loxoprofen dissolution was faster at higher pH. The in vivo pharmacokinetics of loxoprofen was assessed following oral administration of the different loxoprofen formulations to Beagle dogs (n = 22 in total). Secondary peaks or shoulders were observed in many of the individual plasma concentration vs time profiles after ER tablet administration, which may result from secondary absorption in the intestine due to a dissolution rate increase under intestinal pH compared to that observed at stomach pH. In addition, in vivo oral bioavailability was found to decrease with prolonged drug dissolution, indicating site-specific absorption. Based on the in vitro dissolution and in vivo absorption data, a POP-PK IVIVC model was developed using S-ADAPT software. pH-dependent biphasic dissolution kinetics, described using modified Michaelis-Menten kinetics with varying Vmax, and site-specific absorption, modeled using a changeable absorbed fraction parameter, were applied to the POP-PK IVIVC model. To experimentally determine the biphasic dissolution profiles of the ER tablets, another in vitro dissolution test was conducted by switching dissolution medium pH based on an in vivo estimate of gastric emptying time. The model estimated, using linear regression, that in vivo initial maximum dissolution rate (Vmax(0)in vivo) was highly correlated (r2 > 0.998) with in vitro (Vmax(0)in vitro), indicating that in vivo dissolution profiles obtained from POP-PK modeling could be converted to in vitro dissolution profiles and vice versa. Monte Carlo simulations were performed for model validation, and prediction errors for Cmax and AUC were all within the acceptable range (90 to 110%) according to the FDA guidelines. The developed model was successfully applied for the prediction of in vivo pharmacokinetics of a loxoprofen double-layered tablet using the in vitro dissolution profile. In conclusion, a level A IVIVC approach was developed and validated using population modeling that accounted for pH-dependent dissolution and site-specific absorption. Excellent correlations were observed between in vitro and in vivo dissolution profiles. This new approach holds great promise for the establishment of IVIVCs for drug and formulation development where absorption kinetics strongly depend on complex physiologically absorption processes.