Some observations on the effects of time curtailing on the parameters of decaying exponentials fitted to biomedical data

Multiexponential and multicompartmental approaches for analysis of tracer washout data from animal

A comparison is made between multiexponential and multicompartmental analyses of tracer washout data from biological tissues. Various analytical and computer methods are used to explore the relations between kinetic parameters of three-compartment models and parameters of three-exponential functions fitted to typical observed tracer washout records and also to examine relations between kinetic parameters of the same three-compartment models determined analytically and numerically. It is concluded that exponential factors and coefficients cannot always be equated respectively with conpartmental transport rate constants or compartment sizes. Sufficient analytical and computer methodologies exist so that selection of meaningful biological models should not be hampered by a lack of appropriate solution procedures.

The usefulness of 125I-sodium lothalamate as a GFR-indicator in single intravenous injection tests

Clearance of sodium iothalamate was estimated from plasma elimination curves obtained from 90 patients having varying renal function, after a single intravenous injection of 125I-sodium iothalamate. The usefulness of sodium iothalamate as a GFR-indicator with this technique was tested by a strict statistical comparison with conventional inulin clearance as a reference. The regression line, covering the clearance range 2-163 ml min-1, Clinulin = 1.08.Clsodiumiothalamate-3.7, (n = 84, r = 0.85, SEE = 2.3 ml.min-1) was not significantly different from the identity line, which means (i) that extrarenal elimination could not be detected, and (ii) that I125-sodium iothalamate should be regarded as a useful glomerular marker for single intravenous injection studies. An open two-compartment model of mamillary type was found to give an adequate representation of the plasma disappearance curve. The results are not critically dependent on the choice of approximating function. The method by Sapirstein et al., the method by Nosslin and a power-law method were used for comparison and were found to give the same results. However, the present type of curve fitting analyses requires frequent blood sampling or external counting over a long period in order to give reliable estimates of GFR. This circumstance makes these methods less attractive for clinical use.

Weighting functions and data truncation in the fitting of multi-exponential functions

Artificial data were simulated by using two-exponential functions and normally distributed pseudo-random numbers. The variation corresponded to either constant or relative variance for the data error. These data were used to test (i) three different weighting functions and (ii) the effect of data truncation on the precision of estimating the parameters of two-exponential functions.