The effects of the beta-blocker atenolol and nitroglycerin on left ventricular function and geometry in man.

Left ventricular (LV) systolic and diastolic function was studied in 12 subjects with and 12 without coronary artery disease (CAD). Six in each group received the $-blocking agent atenolol, 5 mg i.v., and six received nitroglycerin (NTG), 0.8 mg sublingually. LV cineangiograms were performed in the 400 right anterior oblique projection before and after the drug. Global systolic function was measured as ejection frac- tion and mean normalized systolic ejection rate. Regional function was determined as mean shortening velocity and percentage shortening of basal, middle and apical transverse diameters. Diastolic function was quantitated throughout diastole in terms of logarithmic pressure-volume (log P-V) and logarithmic midwall circumferen- tial stress-midwall circumference (log a-C) relations. Changes in slope k and a of the log P-V and log a-C curves, respectively, were regarded as changes in passive elastic stiffness. LV geometry was analyzed in terms of 1) eccentricity, 2) minor/major axis ratios, 3) shape index and 4) the linear relation between minor/major axis ratio and diastolic volume, with slope q and intercept r. Results varied with the quantity measured and the presence or absence of CAD, but atenolol tended to reduce and NTG to increase parameters of both global and regional systolic function. Velocity indices appeared to be more sensitive than nonvelocity indices in detecting alterations induced by both drugs. With respect to diastolic function, results also varied with the parameter measured and the presence or absence of CAD, but atenolol tended to reduce and NTG to increase passive elastic stiffness of the myocardium. Log a-C relations detected NTG-induced changes when log P-V relations did not. None of the geometric parameters reflected a consistent drug-induced alteration in geometry except r. The data suggest that geometrical influences in drugs may mask alterations in LV diastolic properties and that analysis of log o--C relations may be more appropriate than log P-V relations for detection of these druginduced changes.