In 3 age- and sex-matched groups of subjects--15 normotensives, 15 hypertensives without left ventricular (LV) hypertrophy and 15 hypertensives with LV hypertrophy--the slopes of the regression line obtained by plotting the individual values of LV fractional shortening against the corresponding values of echocardiographic end-systolic stress were compared. The first 2 groups were studied only in control conditions while the third group was restudied after a 20% reduction in LV mass index induced by a long-term antihypertensive treatment and after a 3-week washout period. A significant relation between fractional shortening and end-systolic stress was found in all instances. The slope of this correlation was higher in normotensives (-0.251) and in hypertensives without LV hypertrophy (-0.232) (both p less than 0.01) than in hypertensives with ventricular hypertrophy (-0.079). In this latter group, the slope increased after the reversal of LV hypertrophy (-0.230, p less than 0.01) and remained unchanged (-0.202) at the end of the washout period. No difference was detectable between the slopes obtained in these patients after reversal of LV hypertrophy, both with the antihypertensive treatment on and off, and those of normotensives and hypertensives without LV hypertrophy. Thus, LV hypertrophy attenuates the influence of changes in afterload on LV function. Reversal of LV hypertrophy restores a fractional shortening end-systolic stress relation quite comparable to that found both in normotensives and in hypertensives before the development of LV hypertrophy.