An attempt has been made to test for a reliable method of characterizing the isovolumic left ventricular pressure fall in isolated ejecting hearts by one or two time constants, tau. Alternative nonlinear regression models (three- and four-parametric exponential, logistic, and power function), based upon the common differential law dp(t)/dt = - [p(t)-Ptau]/ tau(t) are compared in isolated ejecting rat, guinea pig, and ferret hearts. Intraventricular pressure fall data are taken from an isovolumic standard interval and from a subinterval of the latter, determined data-dependently by a statistical procedure. Extending the three-parametric exponential fitting function to four-parametric models reduces regression errors by about 20-30 %. No remarkable advantage of a particular four-parametric model over the other was revealed. Enhanced relaxation, induced by isoprenaline, is more sensitively indicated by the asymptotic logistic time constant than by the usual exponential. If early and late parts of the isovolumic pressure fall are discarded by selecting a subinterval of the isovolumic phase, ? remains fairly constant in that central pressure fall region. Physiological considerations point to the logistic model as an advantageous method to cover lusitropic changes by an early and a late tau. Alternatively, identifying a central isovolumic relaxation interval facilitates the calculation of a single ("central") tau; there is no statistical justification in this case to extend the three-parametric exponential further to reduce regression errors., S. F. J. Langer,., and Obsahuje bibliografii