Abstract
We assessed the prognostic value of different indexations of left ventricular (LV) mass in a large cohort of patients with essential hypertension. In a 2-center study, 2180 patients with essential hypertension [mean age 50, 47% women, baseline blood pressure (BP) 160/98 mmHg (SD 13/6)], good-quality echo tracings before treatment and no previous events were followed for an average of 4.6 years (range 0-14). Treatment was tailored to the single patient. During follow-up there were 250 major cardiovascular morbid events, at a rate of 2.47 events per 100 person-years. At entry, LV mass was 107.9 (SD 28) g/body surface area (m2), 118.9 (SD 34) g/height (m) and 49.8 g/height (m2.7). In a Cox analysis stratified by center, after adjustment for age (p<0.0001), 24-hour ambulatory pulse pressure (p=0.0002), current smoking (p<0.0001), total cholesterol (p=0.05) and diabetes (p<0.0001), LV mass provided the best improvement in the model fitting when indexed by body surface area (-2L=2942.2) and by height (-2L=2946.9), while indexation by height2.7 was less efficient (-2L=2949.3; p < 0.05) than the one by body surface area. In a ROC curve analysis, the best efficiency of LV mass for identification of subjects with future events (sens/spec = 1 on the ROC curve) was achieved by the following cut-off points: 111.8 g/body surface area (m2), 122.8 g/height (m) and 51.0 g/height (m2.7). The corresponding prevalence of LV hypertrophy was 37.8%, 39.4% and 40.3%, and the ROC areas were 0.67, 0.66 and 0.67 (p=n.s.), respectively.
