Understanding the High Pharmacokinetic Variability of 5-Aminosalicylic Acid Drug Products Using Compartmental, Semi-mechanistic, and Mechanistic Pharmacokinetic Models.

5-Aminosalicylic acid (5-ASA) is a cornerstone therapy for ulcerative colitis (UC), but its clinical performance is highly variable due to pH-dependent solubility, formulation design, and intestinal metabolism. Significant intersubject variability (ISV), particularly for modified-release drug products such as Lialda (> 75% variability in Cmax and > 50% in AUC) and Pentasa (> 30 to < 50% for Cmax and > 30% for AUC), limits predictable therapeutic outcomes and necessitates a better understanding using biopredictive tools. An empirical, semi-mechanistic, and mechanistic physiologically based pharmacokinetic (PBPK) model was developed using intravenous (IV) and oral data. In vitro dissolution profiles generated under different pH, molarity, and hydrodynamic conditions were fitted and integrated into the PBPK framework to establish biorelevant dissolution methods and predict the in vivo performance of Lialda 1200 mg and Pentasa 4000 mg. Nonlinearity was observed, with 2.7-7.8-fold increases in AUC due to N-acetyltransferase-1 (NAT1) saturation. Optimal predictivity was achieved at pH 7.0 (5 mM) for Lialda (%PE: 4.7% Cmax, 1.2% AUC) and pH 7.5 (25 mM) for Pentasa (%PE: 4.9% Cmax, 4.6% AUC). Variability was driven by NAT1 saturation, pH-dependent solubility/permeability, and formulation effects. PBPK simulations enabled the identification of formulation-specific biorelevant dissolution conditions. The model successfully described pharmacokinetic variability, including intra- and inter-subject differences.