Revisiting Methods for Assessing and Comparing Left Ventricular Diastolic Stiffness

Abstract

Background: Controversies over the pathophysiology of heart failure with preserved EF have renewed interest in characterization and comparison of LV diastolic stiffness, best assessed using the LV end-diastolic (ED) pressure (P)-volume (V) relationship obtained from multiple beats (MB) during rapid preload reduction. As the MB method is difficult, the diastolic P-V over a single beat with P corrected for effect of ongoing relaxation (RCSB) has been used in human studies but not validated against MB method. Further, the LV stiffness coefficient β in the relation EDP = αeβ∗EDV has been used to compare stiffness, but β covaries with α. This covariance is accounted for by comparing derived EDV at similar EDP, calculated from α, β and EDP of 10, 20 or 30 mmHg (EDV10-30). Objective: Determine if the RCSB method of assessing diastolic stiffness (as quantified by β or EDV10-30) shows good correlation and agreement with the MB method. Methods: P, V were invasively measured and the RCSB followed by MB methods performed in 26 young normal or elderly hypertensive dogs before and during phenylephrine (to vary afterload) or acute volume expansion. In the RCSB method, diastolic P was corrected by subtracting the P due to relaxation, monoexponentially extrapolated from isovolumic relaxation. Correlation was assessed by linear regression. Agreement was expressed as the mean ± SD of the paired difference between MB and RCSB values for β or EDV10-30, expressed as percentages of their mean values; p is for paired t-test. Results: Over widely varying afterloads (LV systolic P 95 to 236 mmHg), EDV10-30 derived from RCSB and MB methods showed excellent correlation (r = 0.98, p < 0.0001) and agreement (0.6 ± 5%, p = 0.6). Correlation (r = 0.62, p = 0.002) and agreement (4.2 ± 45%, p = 0.7) between β from SB and MB was poorer. Use of a limited number of P and V points for the RCSB method (at minimum LV P, pre-atrial contraction, and ED) resulted in less agreement between the methods in assessing EDV10-30 (4.4 ± 9%, p = 0.0003) or β (21 ± 40%, p = 0.025). Neither β (12.5 ± 55%, p = 0.5) nor EDV10-30 (10 ± 10%, p < 0.0001) derived from RCSB agreed with MB after marked volume expansion in presence of pericardium. Conclusion: Clinical studies seeking to compare diastolic stiffness between groups can use the clinically more feasible RCSB to substitute for the complex MB technique under a wide range of afterload. Accuracy will be influenced by the number of P and V points used and by presence of acute volume overload. Comparison of stiffness between groups will be enhanced by methods which account for the covariance between α and β.