Analysis of iron kinetics: Identifiability, experiment design, and deterministic interpretations of a stochastic model (Q801844)

The usual stochastic reflux model (SRM) of ferrokinetics is \(V_ f\dot C_ f=Q(C_ f*g-C_ f)+m,\) \(V_ u\dot C_ u=k_ 1QC_ f,\) where \(C_ f\) is the tracer \((^{59}Fe)\) concentration in the plasma, \(C_ u\) is its red-cell concentration, Q is the plasma clearance rate, m is the amount of injected tracer, \(V_ f\) is its initial (plasma) volume of distribution, \(V_ u\) is the red-cell volume, and g is the distribution of extraplasma sojourn times for the fraction of \(^{59}Fe\) which will reflux to the plasma. Four different linear, time invariant 4-compartment models are described as more detailed models reflecting the input-output relations of the SRM. Identifiability of the compartment model parameters is examined and a D- optimal design of ferrokinetic experiments is constructed. The D-optimal ferrokinetic blood-sampling procedure based on the model achieves a 33 \% increase in efficiency over a schedule commonly used.