Background Left atrial pressure (LAP) is often believed to play a dominant role in the determination of left ventricular (LV) early diastolic filling. In a previous study we found no significant relation between LAP and LV early filling velocity (E) but found instead a relation between E and two determinants of LV myocardial shortening (contractility and afterload). To determine if such disparate results may be related to the data ranges of the independent variables in a given population of animals, we took advantage of the differential hemodynamic effects of two modes of volume expansion: saline and whole blood.Methods and Results Eighteen closed-chest anesthetized dogs were instrumented with micromanometers for measurement of LV, left atrial, and aortic pressures. LV volumes were obtained with use of contrast ventriculography, pressures by micromanometry, and transmitral flow-velocity by Doppler echocardiography. After obtaining baseline measurements, group 1 (n=9) received rapid infusion of 500 to 650 mL of saline over 10 minutes, and group 2 (n=9) received the same volume infusion of whole blood. In terms of two known determinants of E, infusion of saline resulted in a significant increase in LAP at the moment of mitral valve opening (X1) (1.5+-0.9 to 5.7+-1.4 mm Hg; P<.05) and a moderate decrease in the pressure decay rate during isovolumic relaxation (tau1/2) (22.9+-2.4 to 26.3+-3.5 milliseconds; P<.05). When these two factors were entered together into a multiple regression analysis with E as the dependent variable, the overall correlation was found to be significant (R=.722; P<.008), with both independent variables contributing significantly to the relation. When factors related to myocardial shortening (afterload and contractility) were added to this relation, they did not significantly improve the prediction of E. Like saline, whole blood infusion augmented X1(1.6+-2.4 to 8.8+-3.2 mm Hg; P<.05) and tau1/2 (21.5+-2.6 to 32.0+-6.3 milliseconds; P<.05) but also significantly increased LV afterload as measured by aortic diastolic pressure (91+-10 to 110+-12 mm Hg; P<.05). Multiple regression analysis of X1and tau1/2 with E again revealed a significant relation (R=.761; P<.002), with both independent variables contributing significantly to the relation. However, in this case, addition of contractility and afterload to the regression significantly improved the relation (R=.909; P<.001), with all four independent variables now contributing significantly to the prediction of E.Conclusions Combined with our previous results, this study indicates the degree to which experimental methods can have an impact on the delineation of the determinants of a phenomenon as complex as LV early diastolic filling. Which independent variables emerge as primary determinants can be strongly influenced by the experimenter's choice of experimental design and manipulations. Specifically, experiments using volume infusion to delineate the responses of the cardiovascular system to variations in loading must allow for the hemodynamic changes that are inherent in the choice of infusate and infusion technique, especially when those interventions may significantly alter blood oxygen-carrying capacity and, in turn, differentially modify factors that affect the magnitude of the early diastolic transmitral pressure gradient. (Circulation. 1994;90:2041-2050.)