Left ventricular hypertrophy has been shown to be an independent predictor of cardiovascular morbidity. Acknowledging the skewed distribution of left ventricular mass, we wanted to develop criteria for left ventricular hypertrophy based on percentiles of left ventricular mass, and observe the effect on estimates of left ventricular hypertrophy prevalences in different subgroups and on the relationship to cardiovascular risk factors in a general population. In a population-based sample of 3287 subjects aged 25-85 years, left ventricular mass was estimated using M-mode echocardiography. A 'healthy' subgroup was used as a reference sample to define sex-specific left ventricular hypertrophy criteria. Sex-specific 97.5 percentiles for left ventricular mass by height, based on the reference sample, were 145.5 and 125.4 g.m(-1), for men and women, respectively. The prevalences of left ventricular hypertrophy in the total population were 14.9% for men and 9.1% for women. The main independent predictors of left ventricular hypertrophy were male gender, body mass index, systolic blood pressure, valvular heart disease, cardiovascular disease and antihypertensive medication. Body mass index and systolic blood pressure had a strong synergistic association with left ventricular hypertrophy in men, but not in women. An alternative framework for defining left ventricular hypertrophy is provided. Body mass index is the culprit factor for risk of left ventricular hypertrophy. Our study indicates that weight reduction is a relevant measure for treatment and possibly prevention of left ventricular hypertrophy in a substantial part of the general population.