Pharmacometric modelling of haemodynamics in humans

Abstract

Current pharmacokinetic-pharmacodynamic models describing changes to the haemodynamic system often do not include necessary feedback mechanisms. These models often provide an adequate empirical description of data but may fail to adequately extrapolate to additional scenarios. This study aimed to develop a minimal model to describe the short-term changes of haemodynamics that can be used as the basis for future model development. The model was used to describe the haemodynamic effects of sodium nitroprusside (SNP) in adolescents undergoing surgery. A minimal haemodynamic model was developed to describe the influence of drugs on blood pressure components. The model structure was defined based on known mechanisms and previously published models. The model parameters were calibrated to describe (without estimation) the heamodynamics of two antihypertensive drugs with data extracted from the literature. Structural identifiability analysis was done using various combinations of the observed variables. The model was applied to clinical data from patients receiving SNP infusion. The model was extended to accommodate the postulated mechanism of action of SNP in the literature. The proposed model structure included mean arterial pressure (MAP), heart rate (HR), and stroke volume and was composed of four states described by differential equations. The model evaluation showed flexibility in describing haemodyanmics at different target perturbations. Overlay plots of model predictions and literature data showed a good description without data fitting. The structural identifiability analysis revealed all model parameters and initial conditions were identifiable only when HR, MAP, and cardiac output were measured together. In addition, model evaluation using SNP data suggested its mechanistic plausibility and the flexibility to describe various response patterns elicited by SNP A minimal model of the haemodynamic system was developed and evaluated. The model accounted for short-term haemodynamic feedback processes. We propose that this model can be used as the basis for future pharmacometric analyses of drugs acting on the haemodynamic system.