Background The efficacy of short-term digitalization on exercise tolerance may, in part, reflect enhanced diastolic performance. However, cardiac glycosides can impair ventricular relaxation from cytosolic Ca++ overload. To detect any time-dependent adverse effect, we assessed the diastolic function after long-term use of digitalis in patients with mild to moderate systolic left ventricular failure. Methods and Results From a cohort of 80 patients who received long-term, randomized, double-blind treatment with digitalis versus placebo at the WJB Dorn Veterans Affairs Medical Center, 38 survivors were evaluated at the end of follow-up (mean 48.4 months) with evaluators blinded to treatment used. Each survivor underwent equilibrium scintigraphic and echocardiographic assessment of diastolic function. Peak and mean filling rates normalized with filling volume (FV), diastolic phase durations normalized with duration of diastole, and filling fractions were measured from the time-activity curve. The isovolumic relaxation period and ventricular dimensions were computed echocardiographically. By actual-treatment-received analysis, treated versus untreated patients manifested a trend toward longer isovolumic relaxation (80.76 ms vs 61.54 ms, P = .06) but a markedly lower peak rapid filling rate (6.39 FV/sec vs 10.56 FV/sec, P = .02) despite comparable loading conditions. In addition, treated patients exhibited a lower mean rate of rapid filling (2.75 FV/sec vs 3.78 FV/sec, P = .05) in the absence of a longer rapid filling duration. However, the end-diastolic ventricular dimension did not differ between the 2 groups. Similar results were obtained by intention-to-treat analysis. Importantly, the mortality rate from worsening heart failure in the inception cohort was lower in the digitalis group versus the placebo group (P = .05) with no difference in total cardiac or all-cause mortality. Conclusions After long-term digitalization for systolic left ventricular failure, cross-sectional comparison with a control group from the same inception cohort shows a decrease in the rate and degree of ventricular relaxation. This effect did not interfere with the overall ventricular filling or with a favorable impact on outcome from worsening heart failure. (Am Heart J 1998;136:688-95.)