BACKGROUND: It is still unknown which factors determine the presence of ventricular late potentials (LP) in hypertension. AIM: To evaluate the prevalence of LP in hypertension in relation to the pattern of left ventricular hypertrophy (LVH) and geometry, and to establish the factors causing signal-averaged ECG abnormalities. METHODS: The study group consisted of 109 patients (58 females, 51 males, mean age 49.7 +/-9.1 years) with hypertension and without coronary artery disease. Two-dimensional Echo Doppler, 24-hr ECG Holter, signal-averaged ECG and spectral analysis of heart rate variability (HRV) were performed. Four patterns of LVH and geometry were identified: normal geometry (N; n=30), concentric remodelling (CR; n=24), concentric hypertrophy (CH; n=38) and eccentric hypertrophy (EH; n=17). RESULTS: LP were more frequently detected in patients with LVH (9.1%), particularly in those with EH, than in patients without LVH (5.6%). Linear regression analysis revealed no correlation between signal-averaged ECG parameters and LV ejection fraction (LVEF) or diastolic LV function indices. None of echocardiographic variables correlated with signalaveraged QRS duration, however, a significant positive correlation between LAS and LV mass (LVM) (r=0.26), LAS and LV end-diastolic volume (EDV) (r=0.2), as well as a significant negative correlation between V40 and LVM (r=-0.22) were noted. A significant positive correlation between LF/HF and signal-averaged QRS (r=0.31) and LAS (r=0.29) as well as a significant negative correlation between LF/HF and V40 (r=-0.21) were found. In the univariate analysis, the presence of EH was significantly related to the occurrence of LP (p<0.01). The reduction of HF power <113 ms(2), indicating a withdrawal of parasympathetic activity, was associated with LP (p<0.05). A ratio of LF 1n/HF 1n >1.28, indicating relative sympathetic overactivity, was a relative risk for LP incidence (p<0.05). In the multivariate analysis, however, all these factors were not independent predictors of the presence of LP. CONCLUSIONS: LP are more frequently detected in hypertensives with LVH, particularly in those with eccentric hypertrophy pattern. Left ventricular structural remodelling and withdrawal of parasympathetic tone are the significant determinants of LP occurrence.