Whether left ventricular mass (LVM) should be normalized to different indexes in relation to body size is still debated. We sought to evaluate the prevalence of left ventricular hypertrophy (LVH) defined by different indexation criteria in a cohort of hypertensive subjects categorized according to body mass index (BMI). A total of 2213 essential hypertensive subjects included in the Evaluation of Target Organ Damage in Hypertension (ETODH) were divided in three groups according to BMI thresholds (<25, 25-29.9 and >or=30 kg m(-2)). All patients underwent extensive investigations including quantitative echocardiography. LVH was defined as an LVM index equal to or higher than (1) 125 g m(-2) in men and 110 g m(-2) in women, (2) 51 g m(-2.7) in men and 47 g m(-2.7) in women. Overall, 687 out of 2213 patients (31.0%) were found to have LVH when LVM was indexed to body surface area (BSA) and 1030 (46.5%) when indexed to height(2.7). A total of 845 patients (38.2%) had normal BMI, 954 patients (43.1%) were overweight and 414 (18.7%) were obese. Prevalence rates of LVH in the three groups were 25.1, 31.6, 41.2% by indexation to BSA and 29.9, 50.5, 71.8% by indexation to height(2.7), respectively. LVM indexed to BSA markedly underestimates LVH prevalence in obese as well as overweight hypertensive patients. To avoid a systematic misclassification of cardiovascular risk, LVM should be routinely indexed to height(2.7) in overweight and obese patients representing a large percentage of the hypertensive population.