Pharmacometric modeling and simulation of ranitidine In human gastric acid secretion.

Abstract

Stress-related gastric mucosal damage commonly occurs in critically ill patients. Prophylactic therapy has become the standard in patients admitted to intensive care unit to prevent the significant morbidity and mortality associated with acute gastric hemorrhage. Ranitidine has been used widely to prevent the gastric mucosal damage in critically ill patients. Data previously collected in two clinical trials included repeated measures of both plasma concentrations of ranitidine as well as multiple gastric acid measurements in critically ill patients and healthy volunteers. These studies were designed to answer a specific research question with a standard hypothesis-testing statistical approach, and those results have been reported. Over the past several years, new research methodologies have been developed or newly applied to analyze this type of data. The aim of this dissertation was to apply new methodologies to develop an appropriate model to simulate the human gastric acid system. A pharmacometric approach places strong statistical and mathematical interpretation to characterize pharmacokinetic and pharmacodynamic behavior and to identify the patient's prognostic factors that cause changes. The pharmacometric analysis showed that 2-compartment pharmacokinetic model, indirect Emax exposure-response response model in sequential approach was superior to all other tested models. Creatinine clearance and sex were identified as important covariates for ranitidine pharmacokinetics. Survival analysis provides unbiased estimates from the time-to-event data with differential follow-ups and censoring issues. A parametric survival approach can then be used to evaluate the model as well as predict observed or simulate future data. It was found that conditional hazard of event increased as time increased. Because it was impossible to discern the thorough exposure-response relationship of ranitidine effect with the range of explored exposure, we concluded that well designed, prospective time-to-event studies were necessary in the future. Simulation implements the physiologically-based mathematical model without invasive and costly experiments. The proposed mathematical model using STELLA® successfully demonstrated physiological gastric acid regulation, the effect of food, and histamine-2 receptor antagonist on gastric acid secretion. The information from this dissertation can be used to determine an optimal dosing regimen of ranitidine that effectively controls gastric pH and to guide future investigations of stress-related gastric mucosal damage.