Arterial compliance measurements using intraarterial pulse contour analysis and a modified Windkessel model were carried out in 19 patients with isolated systolic hypertension (> or = 160/< or = 90 mm Hg) and compared to measurements in 29 patients with essential hypertension (diastolic blood pressure [BP] > or = 95 mm Hg) and 47 normotensive control subjects. Arterial capacitive compliance was significantly lower in isolated systolic hypertension than in essential hypertension (P < .0002) and significantly lower in essential hypertension than in normotensive control subjects (P < .0001). Although the isolated systolic hypertension group was older than the essential hypertension group, the reduction of capacitive compliance in isolated systolic hypertension persisted even when comparison was made with a more nearly age-matched group of essential hypertension. In contrast, oscillatory compliance was reduced similarly in isolated systolic hypertension and essential hypertension compared to normotensive control subjects (P < .0001). Although pulse pressure was greater in isolated systolic hypertension than in essential hypertension, only a weak correlation (r = -0.34) existed between pulse pressure and capacitive compliance. These data indicate that both essential hypertension and isolated systolic hypertension patients exhibit comparably abnormal structure or tone of the small vessels that are the site of oscillations or reflections in the arterial vasculature. In isolated systolic hypertension there is a profound reduction in large artery or capacitive compliance that accounts for the increase in systolic BP and decrease in diastolic BP. This abnormality cannot be accurately assessed by pulse pressure alone.