Load dependence of isovolumic relaxation: facts or artifacts?

In order to study the dependence of left ventricular isovolumic relaxation on preload, afterload, and contractility the effects of infusions of dextran, phenylephrine, and dobutamine were assessed in 10 closed chest anaesthetised dogs. Left ventricular and aortic pressures and left ventricular transverse diameters were measured by micromanometers and a tracking sonomicrometer. Isovolumic relaxation time constant was computed by two different single exponential models: the first (time constant Tw) assumed the horizontal asymptote as equal to zero, whereas the second (time constant Tl) assumed a variable asymptote (Pb). To compare the two models, deviations between observed and predicted left ventricular pressures during isovolumic relaxation were computed for both (average squared difference ARSSQw and ARSSQl respectively). Dextran infusion, although increasing preload indexes, did not affect Tl (from 35.1(2.6) to 38.5(2.2) ms, NS) (mean(SEM], but increased Tw (from 28.4(1.4) to 43.8(2.1) ms, p less than 0.001); Pb was significantly shifted upwards (from -7.9(2.4) to +8.2(2.8) mmHg, p less than 0.01). Pb correlated with left ventricular end diastolic pressure (r = 0.71, p less than 0.001). Phenylephrine infusion did not change the isovolumic relaxation time course (Tl from 36.4(3.5) to 46.2(6.1) ms, NS; Tw from 26.8(2.3) to 30.5(2.9) ms, NS) nor Pb (from -9.5(2.3) to -18.7(2.3) mmHg, NS). Dobutamine infusion reduced Tl significantly (from 35.2(3.7) to 25.3(2) ms, p less than 0.02), but did not change Tw (from 27.5(2.4) to 23.3(3.3) ms, NS) nor Pb (from -7.3(1.8) to -8.8(2.3) mmHg, NS). In 94.85% of the beats considered, ARSSQl was lower than ARSSQw (p<0.001); thus the statistically significant changes observed for Tw after dextran seemed to be due to a poorer fit of isovolumic relaxation data compared with TI rather than to any real dependence of isovolumic relaxation on loading conditions. The fall in isovolumic pressure appears to be more dependent on the inotropic state, being accelerated after dobutamine. Thus the lack of sensitivity of TI to load changes cannot be attributed to a corresponding lack of sensitivity in detecting isovolumic relaxation variations since in addition to a lower load dependence T1 showed a higher sensitivity to inotropic intervention than did Tw. In conclusion, apparent load dependence of isovolumic relaxation observed in the intact heart appears to be critically dependent on methods used for computing the isovolumic relaxation time constant.