Isovolumic relaxation time varies predictably with its time constant and aortic and left atrial pressures: Implications for the noninvasive evaluation of ventricular relaxation

Abstract

The isovolumic relaxation time (IVRT) is an important noninvasive index of left ventricular diastolic function. Despite its widespread use, however, the IVRT has not been related analytically to invasive parameters of ventricular function. Establishing such a relationship would make the IVRT more useful by itself and perhaps allow it to be combined more precisely with other noninvasive parameters of ventricular filling. The purpose of this study was to validate such a quantitative relationship. Assuming isovolumic relaxation to be a monoexponential decay of ventricular pressure (p v) to a zero-pressure asymptote, it was postulated that the time interval from aortic valve closure (when p v = p o) until mitral valve opening (when p v = left atrial pressure, p A) would be given analytically by IVRT = τ[log(p O) − log(p A)], where τ is the time constant of isovolumic relaxation and log is to the base e. To test this hypothesis we analyzed data from six canine experiments in which ventricular preload and afterload were controlled nonpharmacologically. In addition, τ was adjusted with the use of β-adrenergic blockade and calcium infusion, as well as with hypothermia. In each experiment data were collected before and after the surgical formation of mitral stenosis, performed to permit the study of a wide range of left atrial pressures. High-fidelity left atrial, left ventricular, and aortic root pressures were digitized, the IVRT was measured from the aortic dicrotic notch until the left atrioventricular pressure crossover point, and τ was calculated by nonlinear least-squares regression. In 100 cardiac cycles analyzed, IVRT ranged from 21 to 150 msec, systolic blood pressure (p S) from 58 to 180 mm Hg, p O from 37 to 143 mm Hg, p A from 2.5 to 30.5 mm Hg, and τ from 19 to 100 msec. Significant univariate correlations were observed between IVRT and p O ( r = 0.35), p A ( r = −041), and τ ( r = 0.45), as well as with the optimal multilinear combination of all three ( r = 0.92). However, a better ( p < 0.005) correlation was observed when p O, p A, and τ were combined in the proposed equation to yield a predicted IVRT (y) that closely approximated the measured IVRT (x): y = 1.09x + 7.6 ( r = 0.97). A similarly good correlation was observed when systolic blood pressure was used in the expression instead of the aortic closing pressure: y = 1.01x + 23.9 ( r = 0.96). As a secondary purpose of this study, the hemodynamic determinants of the early diastolic transmitral gradient were characterized. It was observed that the initial rate of growth in the atrioventricular pressure gradient ( dΔp dt , x) was almost perfectly correlated with the rate of the decrease in left ventricular pressure at mitral valve opening ( −dp V dt , y): y = 1.01x + 13.7 ( r = 0.993, standard deviation from the regression line [SD reg] = 31.6 mm Hg/sec). Furthermore, −dp V dt (x) was closely associated with the ratio of atrial pressure and τ ( p A τ ): y = 1.11x + 32.8 ( r = 0.90, SD reg = 96.5 mm Hg/sec). In conclusion, the IVRT has a predictable quantitative relationship to τ and to left atrial and aortic pressures. This observation has potential application to the noninvasive characterization of ventricular relaxation.