P-wave dispersion (PWD) is a new electrocardiographic marker that reflects discontinuous and inhomogeneous propagation of sinus impulses, which has been studied in some cardiac conditions as a useful predictor of paroxysmal atrial fibrillation (AF). The aim of the peresent study was to compare P-wave duration and PWD in patients < or =45 versus > or =65 years of age. The study consisted of 2 groups. Group I included 118 patients aged > or =65 years (86 men, 32 women, mean age= 69 +/- 4 years). Group II included 72 patients aged < or =45 years (53 men, 19 women, mean age= 41 +/- 4 years). All patients were selected from those who were undertaken coronary angiography in our hospital with a suspicion of coronary artery disease and detected as having angiographically normal coronary arteries. All patients were undertaken transthoracic echocardiography to evaluate the presence of any structural and functional cardiac abnormality. Maximum and minimum P-wave durations and PWD were calculated from 12-lead surface electrocardiogram. Maximum P-wave duration and PWD were significantly higher in group I patients than in group II patients (P <.001). However, there was no statistically significant difference between group I patients and group II patients regarding minimum P-wave duration (p=0.9). Left atrial diameter, left ventricular wall thicknesses, mitral A velocity, deceleration time and isovolumic relaxation time were significantly higher in group I patients than in group II patients. However, mitral E velocity were significantly lower in group I patients than in group II patients. A significant positive correlation was detected between PWD and age, left atrial diameter, mitral A velocity, deceleration time and isovolumic relaxation time. In addition, we found a significant negative correlation between PWD and mitral E velocity. PWD, indicating increased risk for paroxysmal AF, was found to be significantly higher in patients > or =65 years of age than in those < or =45 years of age. Further prospective studies that include larger series and long term follow-up are needed to clarify the clinical utility of PWD as a predictor of increased risk for paroxysmal AF in old patients.