Quantitation of left ventricular mass and function: balancing evidence with dreams

The quantitative evaluation of the echocardiographic geometry and function for non-ischemic, symmetrically contracting left ventricles is increasingly requested, even when the request is not clinically fully justified and does not take into account the feasibility and reliability of measurements. The general opinion is that, despite a number of technical limitations, the overall information gained from left ventricular (LV) quantitation has a crucial importance for risk stratification, mainly due to the prognostic impact of echocardiographically evaluated LV hypertrophy. This trend tends to automatically transfer epidemiological evidence into clinical application, without consideration of the consequences of the transition of standard errors into single cases. Two recent studies, through differently designed, have demonstrated that the test-retest intraobserver variability of LV mass performs well enough to allow, in most circumstances, the identification of patients with LV hypertrophy. In contrast, the variability of nominal, individual values is high. Tables are available to weight the probability of true biological change when comparing values in the same patient. To a lesser extent, the same conclusions as for LV mass can be applied to measures of systolic function. The technical reliability for measures of diastolic filling is generally good or very good, but the intrapatient variability is probably higher than with measures of LV geometry and systolic function. Moreover, the utility in clinical practice of measures of diastolic filling should be proven. In general, the accurate quantitation of LV geometry and function implies reliable methods and appropriate learning procedures in every echo lab, according to the procedures and the achievements recommended in the current literature. The development of new echocardiographic techniques and the adoption of the procedure of off-line revision of echocardiographic studies might further reduce the variability in the quantitation of measures of LV geometry and function.