Electrocardiography (ECG) has a lower sensitivity for the diagnosis of left ventricular (LV) hypertrophy in smokers than in non-smokers, but the explanation for this finding is not known. In the setting of the Progetto Ipertensione Umbria Monitoraggio Ambulatoriale (PIUMA) study, all subjects smoking > or = 15 cigarettes/day (n = 121, 89 men, age 48 +/- 11 years) were selected from 1443 untreated hypertensive subjects undergoing ECG and M-mode echocardiography, and matched with 484 hypertensive non-smokers by gender (same sex), age (+/- 5 years), and systolic and diastolic blood pressure (both +/- 5 mmHg) in a case-to-control design with a 1:4 matching ratio. Smokers and non-smokers did not differ by age, gender, body mass index, and blood pressure. The voltage of SV1 + RV5 or V6 and RI (p < 0.05), but not of SV3 + RaVL, was lower in smokers. Sensitivity of ECG was lower in smokers when using peripheral or left precordial voltage criteria (e.g. 11 vs 26% for Sokolow-Lyon voltage). When using definitions based on different criteria (voltage of S wave in V3, LV axis, LV strain), sensitivity was not dissimilar in smokers and non-smokers (e.g. 19 vs 18% for Romhilt-Estes score, 40 vs 34% for Perugia criterion). Thus, in hypertensive smokers, sensitivity of ECG is lower than in non-smokers when using peripheral or left precordial voltage criteria, probably due to increased chest size in smokers resulting from increased lung compliance. For LV hypertrophy detection, Sokolow-Lyon voltage should be avoided in hypertensive smokers and replaced by other criteria (Cornell voltage, Romhilt-Estes score, Perugia criterion), which are not influenced by cigarette smoking.