(i). to estimate the yearly incidence rates based on one vs two visits in a working population and (ii). to identify incident hypertension modifiable risk factors. A total of 21566 normotensive subjects were included in a 1-year cohort study. Blood pressure (BP) levels at inclusion and at the second year screening were measured on the basis of two visits, that is, if BP was over 140/90 mmHg in untreated subjects, they were invited to a control visit 1 month later. Height and weight were measured and behavioural risk factors were collected. Among the 17465 subjects who completed the entire protocol (9691 men and 7774 women), 17026 remained normotensive at a 1-year interval and 439 (325 men and 114 women) became hypertensive. Crude yearly incidence rates based on one visit were 6.21% in men and 3.06% in women, compared with 3.04% in men and 1.34% in women when incidence rates were based on two visits, a more than twofold difference. Age and body mass index at baseline were the two major independent determinants of incident hypertension in both genders. Smoking and alcohol consumption were significant risk factors in men but not in women, and a low educational level only in women. BP measurement on separate occasions is necessary to avoid overestimation of incidence. Weight in both genders and alcohol consumption in men were the main modifiable predictors of hypertension.