Gamma variate fits to pharmacokinetic data.

The gamma variate, C = Ataexp(-bt), was tested, as a fitting function, with various real and error-free simulated intravascular and extravascular pharmacokinetic data sets and the results compared with polyexponential fits. For extravascular data, the gamma variate is only suitable to globally fit data which might otherwise be described biexponentially. For intravascular data, the gamma variate could only fit a limited range of the possible concentration-time profiles. Gamma variate fitting algorithms must minimize relative deviations; fits using unweighted sums of squared deviations gave excellent results at higher concentration values but consistently underestimated terminal descending portions of the data.