Objective: To compare the calculations of left ventricular mass according to thick-wall [American Society of Echocardiography (ASE) and Penn convention] and thin-wall (Wikstrand formula) models.
Methods: We have reexamined the data from the cross-sectional study on the general population sample of Vobarno and from a prospective longitudinal study of hypertensive patients assessing the prognostic significance of changes in left ventricular mass during a follow-up period of 10 years on average (Brescia population).
Results: For the Vobarno and Brescia populations, we found a close relationship between values of left ventricular mass calculated by using a thin-wall ellipsoidal model (Wikstrand formula) and those calculated using a thick-wall model with Penn convention or ASE left ventricle measurements (r = 0.99, for both the Vobarno and Brescia populations). Highest values of Penn left ventricle mass were slightly underestimated by use of the thin-wall formula. The numbers of nonfatal cardiovascular events and the relative risks, evaluated by Cox proportional hazard models for 151 patients seen at follow-up did not differ for patients with persistence of left ventricular hypertrophy (LVH), those with regression of LVH, and those with normal left ventricle mass, both at baseline and at follow-up, when these different ways of measuring left ventricle mass and partition values for LVH were used.
Conclusions: The calculation of left ventricle mass according to the ASE recommendations or to the Penn convention, both of which are based on the assumption that the left ventricle can be represented by a prolate ellipsoid with both the internal and external long axes twice the short axis, produces results similar to those obtained using an alternate formula for the calculation of left ventricle mass, considering wall thickness constant around the ellipsoidal cavity. The cardiovascular risk stratification, in relation both to baseline left ventricular mass and to its change during long-term antihypertensive treatment, does not differ significantly among the results of these three different calculations.